0T1. FILE.CQ NSWC TR 86-108 PROTOTYPE RECHARGEABLE UTHIUM BATTERIES BY W. B. EBNER AND H. W. LIN CW' (HONEYWELL POWER SOURCES CENTER) .EDITED AND REVIEWED BY DR. P. H. SMITH AND DR. S. D. JAMES (NSWC) 0') FOR NAVAL SURFACE WARFARE CENTER RESEARCH AND TECHNOLOGY DEPARTMENT I SJUNE 1987 DTIC Approved for public release; distribution is unlimited. ~E L E- TE f SSEP2 71981 NAVAL SURFACE WARFARE CENTER Dahigmn, Virginia 22448-6000 Silver Spring, Maryland 20903-5000 88 9 26 20F
Transcript
0T1. FILE.CQ
NSWC TR 86-108
PROTOTYPE RECHARGEABLE UTHIUM BATTERIES
BY W. B. EBNER AND H. W. LINCW' (HONEYWELL POWER SOURCES CENTER)
.EDITED AND REVIEWED BY DR. P. H. SMITH AND DR. S. D. JAMES (NSWC)
0') FOR NAVAL SURFACE WARFARE CENTERRESEARCH AND TECHNOLOGY DEPARTMENT
ISJUNE 1987
DTICApproved for public release; distribution is unlimited. ~E L E- TE f
SSEP2 71981
NAVAL SURFACE WARFARE CENTERDahigmn, Virginia 22448-6000 Silver Spring, Maryland 20903-5000
6a. NAME OF PERFORMING ORGANIZATION 6b OFFICE SYMBOL 7a. NAME OF MONITORING ORGANIZATION
Honeywell Power Sources Center (Nable) Naval Surfaces Warfare Center (Code R33)
6c. ADDRESS (City, State and Zip Cod) 7b. ADDRESS (City State. an ZIP Cod)
104 Rock Road 10901 New Hampshire Avenue
Horsham, PA 19044 Silver Spring, MD 20903-5000
Ba. NAME OF FUNDING ISPONSORING 8b. OFFICE SYMBOL 9. PROCUREMENT INSTRUMENT IDENTIFICATION NUMBERORGANIZATION 'V aol /raobie)
N60921-84-C-0029
Ec. ADDRESS (City, State, and ZIP Code) 10 SOURCE OF FUNDING NUMBERSPROGRAM PROJECT TASK WORK UNITELEMENT NO. NO. NO. ACCESSION NO.62314N 0 RJI4Y4] 0
II TITLE (InClude S*Curty Caudiatton)
Prototype Rechargeable Lithium Batteries
12 PERSONAL AUTHOR(S)Ebner, W.B. and Lin, H.W.
13a. TYPE OF REPORT 13b. TIME COVERED 14. DATE OF REPORT Year, lonth, Day) S. PAGE COUNTFinal - Phase I FROM 8/84 TO 1/86 1987 June 145
16. SUPPLEMENTARY NOTATION
Edited and reviewed by Dr. P. H. Smith and Dr. S. D. James (NSWC)
17 COSATI CODES 18. SUBJECT TERMS (Continue on reverm if neveuary and idenbfy by bkock niumber)FIELD GROUP SUB-GRQUP RecharI;table Lithium Cells Organic Electrolyte Solutions
07 04 Ester Solvents Low Temperature
Insertion Cathode Materials , High Rate Capabilities
19 ABSTRACT (Conmmue on revern if nfeuay aW idenfti by bkoat number)
This report details the work performed on Phase I of an overall two-phase program.
The aim of Phase I is to select a room-temperature lithium (Li) rechargeable couple
that offers high energy density (60-90 Wh/lb), good rate capability (C/6-C/l), and
low temperature operability.
Among the four cathode systems investigated - V2 05 , TiS2 , V 2 S and LixCoo 2 - the Li/
V 205 couple emerged as the best performer capable of meeting ihe performance goals
described above. The other three cathode systems were dropped from further development
during the course of the program because of their general instability with ester-based
electrolytes, especially the methyl formate (MF)-based electrolyte. The MF-based
electrolyte was selected because of its superior conductivity which is essential for
rate capability and low temperature operability. "... (Cont.)
20. DISTRIBUTIONIAVAILABILITY OF ABSTRACT 21. ABSTRACT SECURITY CLASSIFICATION0-IUNCLASSIFIEOAJNLIMITED (3 SAME AS RPT r- oTIC USERS UNCLASSIFIED
2Za. NAME OF RESPONSIBLE INDIVIDUAL 2b. TELEPHONE (include Area Code) 22c. OFFICE SYMBOLDr. Patricia Smith (202) 394-2948
00 FORM 1473, s4 Mw 63 APR edilton may be used until exhausted. E RITY ASFTION OF THIS AGEAll other editions are obolete.S
UNCLASS IFIED
UNCLASSIFIED
SSCURaV CLA~MPICAlbU OF TXI# PAO6
19. (Cont.)Specifically, the technical accomplishments of Phase I can be stated as follows:
o An electrolyte formulation composed of 2M LiAsF 6 + 0.4 LiBF 4 /methyl formatesaturated with C02 was successfully developed with the following demonstratedcapabilities:
- High Conductivity - 43 mmho/cm at room temperature and 13 mmho/cm at -400 C.This solution at -40 0 C is still three times more conductive than the morecommonly used ether-based solution (e.g., LiAsF 6 /2-methyl tetrahydrofuran (THF))at ambient temperature.
- 93% Lithium Cycling Efficiency - this lithium efficiency enables the use ofmethyl formate-based solution in practical hardware. Li/V205 cell cyclabilitvup to 50 cycles can be projected at the 100% depth of discharge based onLi/V 2C5 r-tio of 3:.
o Li/V 205 cyclability surpassing 100 cycles in laboratory cells - cycled cell capacity
can be held at a level corresponding to 80% cathode discharge efficiency. Theability to achieve this high level of cathode utilization is due to our focuseddevelopment of the cathode processing technology. Particle size of the V2 05 wasfound to be key to enhanced cell performance and electrode integrity was criticalto achieving cvclabilitv without degrading cell capacity.
Li/V 2 05 laboratory cell produces 366 Wh/kg based on lithium and active cathodematerial (or 166 Wh/lb) at the 100th cycle when tested at room temperature. Usingthis value, practical energy densities of 50 Wh/lb and 60 Wh/lb can be projectedfor "D" cell and No. 6 size cell, respectively.
o Discharge current density surpassing 5 mA/cm 2 - at average 55% cathode discharge
efficiencv for 205 cycles - this allows addressing applipation of the T U/V 2()5technology in missions requiring rate capability up to C/1.3.
Phase I of this program, therefore, successfully selected a rechargeable technology(Li/V 2 05 ) that is ready for transitioning to a hardware development phase. Phase Idemonstrated overall performance capabilities of the Li/V 2 05 technology at thelaboratory cell level, and the objective of Phase II is to further improve anddemonstrate the stated capabilities of Li/V 2 05 cell at the hardware level via a 30 Ah
cell. Aooession For
NTIS GRk&I
DTIC TAB
Un ouncod 0Jut If tent 1,;n
LDlstribtlom/___ ......
NSP Availabillty Codes
;Av ail aad/or
UNCLASSIFIED
SECURITY CLASSIFICATION OF TIS PAGE
NSWC TR 86-108
FOREWORD
The development of a reliable rechargeable lithium battery technology
having high rate capabilities and low temperature operability has become amajor priority throughout the Department of Defense. In recent years.considerable progress has been made in ambient temperature lithium/secondary
technology. However, this technology has not produced a practical Li/secondary
cell nor have high rate or low temperature performance capabilities beendemonstrated. To meet the demanding performance goals for naval applications.such as torpedo targets and seal delivery vehicles, significant advances are
needed over the current state-of-the-art capabilities. To achieve theseadvances. Honeywell under the direction of the Electrochemistry Branch, R33. is
developing an ester-based rechargeable technology that has the capability for
both high rate and low temperature operations.
The authors acknowledge the assistance of P. Lensi in carrying out the
cathode material syntheses, Dr. Wayne Worrell of The University of Pennsylvania
for his consultation on cathode material structures, and Mrs. S.P.S. Yen of the
Jet Propulsion Laboratory for her helpful discussions on TiS The authorsare also indebted to Dr. P. H. Smith and Dr. S.D. James of NiWC for theirtechnical guidance, program direction, and continued support throughout thecourse of this work. This work was sponsored by the Office of Naval Technology
under the Mine and Special Warfare Technology Block NS3B.
Approved by:
CARL E. MUELLER, Head
Materials Division
iii/iv
NSWC TR 86-108
CONTENTS
Chapter Page
I INTRODUCTION..................................................... 1-1
3-24 X-RAY DIFFRACTION PATTERNS FOR TiS CATHODES BEFORE AND AFTER
EXTENDED CYCLING IN A 2M LiAsF + 8.4M LiBF /MF ELECTROLYTESOLUTION (CHARGE/DISCHARGE RAT§ = 1.0 mA/cm AT AMBIENT
TEMPERATURE (20-24'C)) ............................................... 3-40
3-25 CHARGE/DISCHARGE BEHAVIOR OF A Li/TiS CELL WHEN CHARGED TO3 VOLTS EMPLOYING A 1.5M LiAsF 6 /MA ELiCTROLYTE SOLUTION .............. 3-41
3-26 TiS SCREENING STUDIES: CYCLE LIFE PERFORMANCE IN METHYLACETATE SOLUTIONS .................................................... 3-42
3-27 TiS SCREENING STUDIES: CYCLE LIFE PERFORMANCE IN METHYLACETATE SOLUTIONS AS A FUNCTION OF LiAsF6 CONCENTRATION..............3-44
3-28 TiS SCREENING STUDIES: CYCLE LIFE PERFORMANCE IN 3M LiAsF 6 /MASOLUTIONS EMPLOYING EXTENDED VOLTAGE LIMITS ........................... 3-45
viii
-- . .. ---.. -- '---- w-- - -_
NSWC TR 86-108
ILLUSTRATIONS (CONT.)
Figure Page
3-29 TiS SCREENING STUDIES: EXTENDED CYCLE LIFE PERFORMANCE WITH
3M L AsF 6 /MA SOLUTIONS ............................................... 3-46
3-30 TiS SCREENING STUDIES: CYCLE LIFE PERFORMANCE IN 2M LiAsF /MA
SOLUTIONS WITH AND WITHOUT 1,2-DIMETHOXYETHANE (DME) CO-SOLhENT ...... 3-47
3-31 TiS SCREENING STUDIES: CYCLE LIFE PERFORMANCE IN 2M LiAsF /MASOLUTIONS WITH AND WITHOUT TETRAHYDROFURAN (THF) CO-SOLVENT. ......... 3-48
3-32 TiS SCREENING STUDIES: CYCLE LIFE PERFORMANCE WITH AND WITHOUT2-METHYL THF CO-SOLVENT .............................................. 3-49
3-33 Li Coo SCREENING STUDIES: CYCLE LIFE PERFORMANCE WJTH 2MLi~sF6 MA SOLUTIONS AT A DISCHARGE RATE OF 1.0 mA/cm ............... 3
3-34 Li CoO SCREENING STUDIES: CYCLE LIFE PERFORMANCE WITH 2M liAsF6/MASOUTIONS AT A DISCHARGE RATE OF 5 mA/cm............................ 3-52
3-35 LITHIUM CYCLABILITY TESTS: TYPICAL VOLTAGE PROFILES FOR METHYL FORMATESOLUTIONS IN WICK CELLS (THIRD CYCLE) ................................. 3-68
3-36 LITHIUM CYCLABILITY TESTS: TYPICAL VOLTAGE PROFILES FOR METHYL FORMATESOLUTIONS IN WICK CELLS (TENTH CYCLE) ................................ 3-69
3-37 LITHIUM CYCLABILITY TESTS: IR SPECTRUM OF SOLID PRODUCT FROM 2MLiAsF6 /MF CELL ....................................................... 3-70
3-38 LITHIUM CYCLABILITY TESTS: IR SPECTRUM OF SOLID PRODUCT FROM 2MLiAsF6 + 0.4M LiBF 4/MF CELL .......................................... 3-71
3-39 LITHIUM CYCLING EFFICIENCY VERSUS LiAsF 6 CONCENTRATION FOR METHYLFORMATE SOLUTIONS IN WICK CELLS ...................................... 3-73
3-40 LITHIUM CYCLING EFFICIENCIES FOR XM LiAsF + X/5M LiBF /MF SOLUTIONSWITH AND WITHOUT 002 PRETREATMENT OF THE ORKING ELECTrODES .......... 3-77
3-41 LITHIUM CYCLABILITY TESTS: TYPICAL VOLTAGE PROFILES FOR 2M LiAsF6
+ 0.4M LiBF/MF SOLUTIONS IN WICK CELLS WITH AND WITHOUT C 2 .ADDED TO SOLUTION (THIRD CYCLE) ..................................... 3-78
3-42 LITHIUM CYCLABILITY TESTS: TYPICAL VOLTAGE PROFILES FOR 2M LiAsF6+ O.4M LiBF 4 /MF SOLUTIONS IN WICK CELLS WITH AND WITHOUT C02ADDED TO SOLUTION (TENTH CYCLE) ...................................... 3-79
3-43 IR SPECTRUM OF SOLID REACTION PRODUCT FORMED ON LITHIUM CYCLED IN
3-44 LITHIUM CYCLING EFFICIENCY VERSUS LiAsF 6 CONCENTRATION FOR METHYLACETATE SOLUTIONS .................................................... 3-84
3-45 IR SPECTRUM OF SOLID REACTION PRODUCT FORMED ON LITHIUM CYCLEDIN IM LiAsF 6 /DMSI ............................... .................... 3-87
3-46 LITHIUM CYCLING EFFICIENCY VERSUS LiAsF 6 CONCENTRATION FOR DIMETHYLSULFITE SOLUTIONS (HALF-CELL RESULTS) ................................ 3-90
3-47 Li/V 0 CELL CYCLE LIFE PERFORMANCE FOR VARIOUS CATHODEMANUFAUTURING PROCESSES .............................................. 3-94
3-48 EFFECTS OF CARBON TYPE ON THE PERFORMANCE OF ROLLMILLED V2 0 5 CATHODES ................................................. 3-98
3-49 EFFECTS OF TEFLON BINDER CONTENT ON THE PERFORMANCE OF THE ROLL MILLEDV205 CATHODES ........................................................ 3-100
3-50 EFFECTS OF CARBON CONTENT ON THE PERFORMANCE OF ROLL MILLEDV205 CATHODES ........................................................ 3-101
3-51 PARTICLE SIZE DISTRIBUTIONS FOR THE METALLURG V2 05 VERSUSTHAT FOR V^O 5 MANUFAc.IURED FROM GROUND AMMONIUMMETAVANADATE (LOT CML-V2-003) ........................................ 3-102
3-52 EFFECTS OF V20 5 PARTICLE SIZE ON THE PERFORMANCEOF ROLL MILLED CATHODES .............................................. 3-103
3-53 DEMONSTRATION OF THE EXTENDED CYCLE LIFE CAPABILITIES OF THERECHARGEABLE Li/V2 05 SYSTEM .......................................... 3-105
3-54 EXTENDED CYCLE .IFE PERFORMANCE OF L '/V 2 .5 SYSTEM AT THE DISCHARGERATE OF 5 mA/cm........................................................ 3-106
x
NSWC TR 86-108
TABLES
Table Page
1-1 SUMMARY OF MATERIALS TO BE INVESTIGATED ............................... 1-2
2-1 METHODOLOGY OF CATHODE MATERIAL CHARACTERIZATIONS .................... 2-6
2-2 CATHODE DESCRIPTION FOR V20 5 SCREENING STUDIES ....................... 2-12
2-3 TEST PARAMETERS FOR V20 5 CATHODE SCREENING STUDIES ................... 2-12
3-1 SUMMARY OF SCREENING EVALUATIONS CONDUCTED IN THIS PROGRAM ........... 3-2
3-2 X-RAY DIFFRACTION DATA FOR V2 05 (LOT CML-V2-001) ..................... 3-7
3-3 X-RAY DIFFRACTION DATA FOR V2 05 (LOT CML-V2-002) ..................... 3-8
3-4 X-RAY DIFFRACTION DATA FOR V2 05 (LOT CML-V2-003) ..................... 3-9
3-5 X-RAY DIFFRACTION DATA FOR DEGUSSA TiS2 (AS-RECEIVED) ................ 3-14
3-6 X-RAY DIFFRACTION DATA FOR LixCoo 2 (LOT CML-CO-002) .................. 3-22
3-7 X-RAY DIFFRACTION RESULTS FOR Li Coo 2 (LOT CML-CO-003) ............... 3-23
3-8 DISCHARGE RESULTS FOR Li/Li xCoo2 CELLS ................................ 3-29
3-9 AMPUL STORAGE TEST RESULTS FOR 2M LiAsF6 + 0.4M LiBF 4 /MF SOLUTIONS (HISTORICAL DATA) ....................................... 3-53
3-10 AMPUL STORAGE TEST RESULTS FOR 1.OM LiAsF 6/DMSISOLUTIONS (HISTORICAL DATA) .......................................... 3-54
3-11 V 0 THERMAL STABILITY TESTS--PHYSICAL APPEARANCE OF SAMPLESAiTER STORAGE INTERVAL ............................................... 3-55
3-12 SOLUBILITY OF V2 05 IN LITHIUM BATTERY ELECTROLYTES--BY AA ANALYSIS...3-55
3-13 X-RAY DIFFRACTION DATA FOR V 0 STORED IN 2M LiAsF6 + 0.4M LiBF4 /MF AT AMBIENT TEMPERATURE (2- 4°C) .................................. 3-57
3-14 X-RAY DIFFRACTION DATA FOR VO STORED IN 2M LiAsF 6 /MA ATAMBIENT TEMPERATURE (20-240Cd. ......................................... 3-58
3-15 X-RAY DIFFRACTION DATA FOR V205 STORED IN 2M LiAsF6 + 0.4MLiBF4/MF AT 710C ..................................................... 3-59
3-16 X-RAY DIFFRACTION DATA FOR V205 STORED IN 2M LiAsF 6/MA AT 71'C ....... 3-60
3-36 ROLL MILLED V 0 CATHODE DESCRIPTIONS AND CYCLEPERFORMANCE S ARY ................................................. 3-96
3-37 PERFORMANCL SUMMARY FOR Li/V 0 LABORATORY CELLS TESTEDUNDER EXTENDED CYCLE LIFE CONDITIONS ................................. 3-107
xiii/xiv
NSWC TR 86-108
CHAPTER 1
INTRODUCTION
The development of a reliable rechargeable lithium (Li) technology havinghigh rate capabilities and low temperature operability has become a majorpriority throughout the Department of Defense (DOD). In recent years,considerable progress has been made in ambient temperature Li/secondary
technology. At low rates of discharge, good cycle performance has beendemonstrated with a number of cathode materials. However, this technology has
not produced a practical Li/secondary cell nor have high rate or lowtemperature performance capabilities been demonstrated. To meet the demandingperformance goals for naval applications, such as torpedo targets and sealdelivery vehicles, significant advances are needed over the currentstate-of-the-art capabilities. To achieve these advances, Honeywell under thedirection of NSWC investigators offers the development of an ester-basedrechargeable technology that has the capability for both high rate and low
temperature operations.
The objective of this program, therefore, is to develop a lithiumrechargeable cell technology capable of operating over a wide temperature rangeand offering good discharge rate capabilities. The specific program
performance guidelines are:
o Energy Density: 60 to 90 Wh/lbo Depth of Discharge: 75 percento Rate Capabilities: C/6 to 2Co Cycle Life: 50 to 100 cycles at C/6 to 2C rateso Operating Temperature Range: -55*C to 751C (generic applications)
-20C to 350 C (seawater applications)
o Storage Temperature Range: -550 C to 750C
o Safety: Safe and resistant to abuse.
The approach adopted to achieve these performance capabilities is to employinsertion-type cathode materials (vanadium pentoxide [V2o5 ], titaniumdisulfide [TiS ], vanadium V) sulfide [V2S 5 ], and lithium cobalt oxide[Li Co02]) witi high conductivity, ester-Eased electrolyte solutions.Insertion cathode materials were selected based on their demonstrated longcycle life capabilities and excellent reversibility in various nonaqueouselectrolyte systems. Ester-based solutions offer some of the highest
conductivities achievable in organic solutions thus providing the intrinsic
i-1
NSWC TR 86-108
properties needed to significantly enhance rate capabilities and lowtemperature operation. These solutions also exhibit good thermal, chemical,and electrochemical stability, as evidenced by their proven performance inprimary lithium applications.
The present program is divided into two parts. Phase I involves selectionof the electrochemical system offering the best performance capabilities, whilePhase II will involve the development and testing of prototype hardware cellsincorporating the selected technology. This report details the work conductedin Phase I.
The specific cathode materials and electrolyte solutions that have beeninvestigated are summarized in Table 1-1. The superior conductivity offered bythe ester-based solutions over the propylene carbonate- and 2-methyl THF-basedsolutions used in the current state-of-the-art rechargeable lithium cells isclearly illustrated in Figure 1-1. Figures 1-2 and 1-3 show the dramaticenhancements in rate capabilities that can be achieved with the ester-basedsolutions. Of the four cathode materials evaluated, V 0 and TiS2 aremature systems and therefore represented our primary cinaidates. V S and
2 5Li CoO2 , on the other hand, are new systems that could offer significantadvantages in energy density if extended cycle life capabilities could bedemonstrated.
TABLE 1-1. SUMMARY OF MATERIALS TO BE INVESTIGATED
The major challenge at the outset of the program was to demonstrate cyclelife capabilities with the ester-based solutions including both reversibility ofthe insertion cathode materials and efficient cycling of the lithium electrode.in addition, however, it was recognized at the start that cathode processing wasa major factor in the successful development of a practical rechargeable celltechnology and. therefore. cathode processing has been emphasized throughout ourwork.
In Phase I. we have successfully identified an electrochemical systempossessing the intrinsic properties to greatly extend the state-of-the-artcapabilities in rechargeable lithium technology and to meet the demandingperformance goals for naval applications. This system consists of V205 asthe cathode material, methyl formate electrolyte solutions, and roll milledcathodes. Carbon dioxide (CO 2) will be used as an electrolyte additive toprovide high lithium cycling efficiencies in the methyl formate solutions.
1-2
NSW'C TR 86-108
08t
01-4
-~ 0L
'- '-
P4 0D04
H
w z
.0~
-&3 HO 1'-T rE
~ H ~L.*0>
Uzf H
ci~C-4 en flM
00 - Ls -1. en
MZ/Q1BKM :aONJVflQMOD DMaloadS
1-3
NSWC TR 86-108
00*001
;> E
E--
-im 0
000
'-4~ E-i~
E-
E-z E
W wU
0-0
CC
.. . . 00 0
CWC
1-4
NSWC TR 86-108
U E
CX
001T
08
(N C4CN4 cn z
C 4 4 z
09' -
+ L
(Ni> wii C4
~ 0
CN. C-4 -44 Q
CJ4 E-4
00*0
SI'LOA :'IIN310d 1130
1-5
NSWC TR 86-108
CHAPTER 2
EXPERIMENTAL
EXPERIMENTAL CELL DESCRIPTIONS
Half-Cell
The half-cell configuration, shown in Figure 2-1. incorporates threeelectrodes including a reference electrode. The anode and cathode each had ageometric surface area of 3.2 cm2 . In cycle life tests, the anodes weresealed in a Celgard 2400 envelope while the cathodes were supported with porousglass disks to maintain good electrical contact with the expanded metal currentcollectors. Once assembled, the electrode holder was completely immersed inapproximately 15 ml of electrolyte solution contained in a Fisher & Porter3-ounce aerosol reaction vessel. The leads were brought out through Tefloncompression seals mounted in the lid of the reaction vessel.
Wick Cell
The wick cells consisted of an electrode stack mounted between two glassslides held together with stainless steel wire and sealed in a Fisher & Porter3-ounce aerosol reaction vessel. These cells employed approximately 1.5 ml ofsolution in the bottom of the reaction vessel with only the ends of theseparator immersed in the electrolyte solution. Transport of the electrolytesolution to the electrode surfaces occurred through the wicking action of theseparator. Two configurations of wick cells were employed in ourinvestigations. The 2-plate cells, illustrated in Figure 2-2, consisted of oneanode and one cathode separated by porous separator material. The 3-platedesign incorporated two anodes connected in parallel, one on either side of thecathode. The separators used were either Celgard 2400 microporouspolypropylene or Hollingsworth and Vose HV-932 glass fiber material.
CATHODE MANUFACTURING TECHNIQUES
Six cathode processes were evaluated in this program. A brief descriptionof each process is given below.
2-1
NSWC TR 86-108
400
ad 0 '
c (U
0 0
2-24
NSWC TR 86-108
'E-4
C.3,
W4 -4
2-30
NSWC TR 86-108
Cold Pressing
Pressed powder electrodes were made by mixing dry powders of the activecathode material and conductive diluent in a high-speed blender and thenpressing the resulting mixture onto both sides of an expanded metal grid at6000 to 8000 psi. The thickness was controlled by the weight of cathode mix.
Cold Pressing with Teflon Binder
These cathodes, containing active cathode material, conductive diluent andTeflon-6 powder (DuPont). were made in the same manner as described above.
Cold Pressing With Microthene (Polyethylene)
These cathodes were made in the same manner as described above, except thatthe pressed electrodes are sintered under argon atmosphere at 100°C for onehour.
Cold Pressing and Sintering with Teflon Binder
These cathodes were made in the same manner as described above except thatthe pressed electrodes are sintered under vacuum at 280°C to 3000C for oneto two hours.
Cold Pressing with Isotactic-Polypropylene Binder1
This cathode processing technique was developed by U.S. Army Laboratory
Command (LABCOM) and involves dissolving isotactic-polypropylene powder nearits crystalline melting point (100 0 C to 1200 C) in a small volume of Decalin(decahydronaphthalene). The cathode material and conductive diluent are thenadded with vigorous stirring. Stirring is continued until the solution hascooled to room temperature. The resulting mixture is dried under vacuum at150 C to remove the residual solvent and then ground to a fine powder in ahigh-speed grinder. Cathodes are then made by pressing the powder onto bothsides of an expanded metal grid.
Pasting with Ethylene Propylene Diene Terpolymer (EPDM) Binder
This process was developed at the Jet Propulsion Laboratory for use withTiS 2 cathodes. Electrodes are prepared by mixing dry powders of cathodematerial and conductive diluent in a high-speed blender for 30 seconds.Cyclohexane containing one weight percent dissolved EPDM is then added insufficient quantity to provide the required amount of binder. After stirring.the excess solvent is removed under vacuum until the mixture has achieved apaste-like consistency. The wet mixture is then pasted onto both sides of anexpanded metal grid. The finished electrodes are vacuum dried at ambienttemperature for several hours and then at 800C for one hour.
2-4
NSWC TR 86-108
Roll Milling with Telfon Binder1
In this process, the cathode mix was prepared by first blending Teflon-6
powder in the carrier solvent (mineral spirits) for 20 to 30 seconds using a
high-speed blender. Next, the cathode material and conductive diluent were
slowly added to the solution and blended for one minute. The mix was then
vacuum filtered and kneaded in a polyethylene bag until it obtained a clay-like
consistency. The material was then passed through a rolling mill approximately
12 times, being rotated 90 degrees after every two passes. This process
produces a flexible sheet which was then air dried overnight and then vacuum
dried for approximately three hours at 2000 C. Cathodes were made by cutting
out electrodes to the desired dimensions and pressing them onto expanded metal
grids.
CATHODE MATERIAL SYNTHESIS AND CHARACTERIZATION
Of the four cathode materials evaluated. V2 05 . V2 S5 . and LixCoOv
were synthesized in-house while TiS 2 was purchased from a commercia vendor
(Degussa, battery grade). Before use, all cathode materials were characterized
using the following series of analyses:
o Stoichiometryo Particle Size Distribution
o Purityo Morphology
o Crystal Structure
o Electrochemical Discharge Performance
The methodology for these analyses are summarized in Table 2-1.
Vanadium Pentoxide (V2 05 )
Vanadium pentoxide was manufactured by the thermal decomposition of
ammonium metavanadate (2NH4 VO 3 ) at 3800 C under flowing air per the
electrolyte solution at ambienttemperature (20-24 0C)
EDAX = Energy Dispersive AnalysisSEM = Scanning Ele~tron Microscope
2-6
NSWC TR 86-108
uniform layer 0.6 to 0.8 inch thick on the inside wall of the glass tube. Thetube was then slowly heated to 380 0C and maintained at that temperature for aperiod of three hours employing a dry compressed air flow rate of 10 liters/minute. The product was then annealed at 5600C for 30 minutes at an air flowrate of 5 liters/minute after which the tube was slowly cooled to roomtemperature, maintaining the air flow at 5 liters/minute.
During the course of our evaluations, three lots of V O were manufactured,designated by lot numbers CML-V2-001 to CML-V2-003. Lots 01 and 002 employedas-received ammonium metavanadate while Lot 003 employed ammonium metavanadatewhich had been ground in a ball mill in an effort to reduce the particle sizeof the manufactured V205.
Titanium Disulfide (TiS2)
Extensive work has been done with titanium disulfide by other investigators,the results of which show that cycle life performane is strongly influenced bythe stoichiometry of the starting cathode material. 2- 4 If the TiS ismetal rich, the excess metal atoms can occupy sites in the van der 2Waals gapeffectively pinning the layers together. This can greatly impede the mobilityof lithium ions and degrade rate capabilities. It also restricts the amount oflithium that can be incorporated, thus lowering capacity as well. Excesssulfur, on the other hand, can dissolve in the electrolyte solution andsubsequently react with the lithium metal anode resulting in reduced lithiumcycling efficiencies and increased polarization levels.
TiS 2 is costly and difficult to produce. At the outset of the program,we were convinced that without a commercial source of supply, TiS 2 wouldnever be a practical system, regardless of how well it performed. Therefore.our approach was to purchase the best available grade of TiS 2 for ourevaluations rather than synthesize it in-house. In this way, we could beconfident of a rapid development cycle into practical hardware if TiSemerged as the optimum cathode material for use with eater-based soluions.
. The decision to purchase TiS2 was strongly influenced by the emergence ofa new vendor, offering a high quality grade of material. This vendor isDegussa. who had earlier been selected by Exxon to be a large scale producer ofTiS 2 for their lithium/battery program. Exxon pioneered much of the earlydevelopment of the TiS2 system and established the manufacturing methodi andmaterial specifications needed to achieve a high quality grade of TiS2.-Because of their ties with Exxon, we were confident that Degussa wasknowledgeable in the material requirements for rechargeable lithiumapplications and could supply a high grade of material for our evaluations.
2-7
NSWC TR 86-108
Vanadium (V) Sulfide (V2s5)
V2S was prepared in a two-step process involving the followingreactio~s:
Ammonium thiovanadate ((NH4 )3 VS4 ) was prepared by reacting ammoniummetavanadate (NH vo ) with hydrogen sulfide (H2S) in a solution ofammonium hydroxie %NH OH). This was accomplished by slowly bubbling H2Sthrough a saturated sotution of NH4VO in NH4OH with continuousstirring. The reaction flask was immhrsed in an ice-water bath to dissipatethe heat from the exothermic reaction. Addition of H2S was continued for 10hours after which the fine brown precipitate that had formed was collected byfiltration under a nitrogen blanket and washed with ethanol. The filter cakewas vacuum dried at ambient temperature for 48 hours, crushed with a mortar andpestle, and then redried for another two hours under vacuum at ambienttemperature.
V S was then prepared by the thermal decomposition of (NH4)3VS4in a iuge furnace under argon flow. Two lots of material were preparedemploying different temperatures to evaluate the effects of decompositiontemperature on product quality.
The first lot of V 2S was prepared by slowly heating the(NH4) VS4 to 240°C and holding it at that temperature for one hour.The tmperature was then increased to 2500C for one hour and then to 2600 Cfor 40 minutes. The tube was then allowed to cool to ambient temperature andthe product collected under an argon atmosphere. This lot was designated asCML-VS-002.
The second lot of V S was prepared by slowly heating the(NH )3V5S to 225 0C and Kaintaining it at that temperature for 3.5hou~s. The tube was then allowed to cool to ambient temperature and theproduct collected under an argon atmosphere. This lot was designated asCML-VS-003.
Lithium Cobalt Oxide (Li Coo 2)
Li CoO 2 was prepared by the thermal decomposition of a pelletizedmixtur of lithium carbonate (Li CO ) and cobaltous carbonate (CoCO3 ) at9000C in air. The synthesis wai carried out by first mixing Li2 C03and CoCO 3 together in a Li/Co mole ratio of 0.98 and then pressing t'hemixture into 1.55-inch diameter by 0.20-inch thick pellets at 10,000 psiemploying a dwell time of one minute. The pellets were then heated at 9000C
in a muffle furnace for 20 hours under an air flow rate of 1.8 L/minute. Aftercooling to room temperature, the pellets were ground in a mortar and pestle andtwice reheated to 9000 C for 20 hours under an air flow rate of approximately0.6 liters/minute. Two lots of material were prepared in this manner,designated as CML-CO-002 and CML--00-003.
2-8
NSWC IR 86-108
ELECTROLYTE SOLUTIONS
Purification
The general procedure employed for purifying ester solvents is to distillthem from P205 . This method was developed in the mid-1970's by Honeywellfor purifying met'yl formate and has been found to be extremely effective forremoving water and methanol, the primary impurities found in commercial gradesof methyl formate. E.M. Science (formerly MC/B) now uses this method tomanufacture a special grade of methyl formate specifically for Honeywell foruse in our lithium batteries.
Analysis of solvents was accomplished employing the Karl Fischer method forwater and gas chromatography for other volatile impurities. Solutions wereanalyzed for water using tl'. Karl Fischer technique.
Methyl Formate. Methyl formate was purchased from E.M. Science (specialbattery grade) and used as received. The only impurity detected in thepurchased material was water at a concentration of 22 ppm.
Methyl Acetate. Methyl acetate was purchased from E.M. Science (Cat. No.MX0625; 99+ mol percent) and was purified by distillation from P205.Interestingly, the P205 turned black after a short period of contact withmethyl aretate. The reason for this color change is not known, but nodegradation of the solvent or other adverse effects were observed, even afterprolongpd treatment with P 0 5 As with methyl formate, d4stillation fromP205 was found to be an efiective means of purifying methyl acetate.
Dimethyl Sulfite. Dimethyl sulfite (Eastman Kodak Cat. No. 5874) was foundto form a gel with P 05 and thus could not be purified in the same manneras methyl formate and methyl acetate. Instead, this solvent was purifiedemploying a double fractional distillation under reduced pressure.
Solution Manufacture
All solutions were manufactured in a glove 'ox under an argon atmosphere.The purified solvents were prechilled to -100 C to -200 C to minimize heatbuildup during the exothermic dissolution process. The salts used were:
LiAsF6 kLithium hexafluoroarsenate): Electrochemical Grade, USS Agri-chemical -- used as received.
LiBF, (Lithium tetrafluorobromate): Foote Mineral Co. -- used as received.
All solutions contained less than 50 ppm of water, as measured by the KarlFischer method.
2-9
NSWC TR 86-108
LITHIUM CYCLING EFFICIENCY MEASUREMENTS
The lithium cycling efficiency tests were conducted by stripping and
plating lithium between two lithium electrodes. This arrangement is designed
to simul: e the conditions in an actual cell under charge/discharge cycling andhas bee:. found to provide more reliable results than stripping and plating onto
an inert -,bstrate.
The worki,g electrode was capacity limiting and consisted of 0.0035-inch
--hick lithium foil pressed onto a nickel o- 304 stainless steel substrate with
an 2-jan,!ed metal grid spot welded to it. The auxiliary electrode employed
'.020-inch thick lithium foil prIssed onto nickel grid. Each electrode had a
geometric surface area of 3.2 cm . Tests were conducted with both 2-platewick cells (Figuie 2-2) and flooded half-cells (Figure 2-1). In the wick cell
tests, 2-5 layers of Celgard 2400 were employed as the separator.
The t~sts were conducted euploying a stripping/plating current density of
1.0 mA/cm' and cycling was continu-d unt;l the lithium working electrode wascompletely consumed. Each ha'P- ce w.- t1ree hours in duration which
represented approximately 20 percent depth discharge based on the starting
capacity of the working electrode. TWi cycling -fficiency was determined using
the following re'lationship:
C - C.
C. - 0 1
nE x 100 (2-4)
C.1
where:
E = Cycling efficiency, percent
C = Starting capacity of lithium working electrode. mAh
C i = Capacity of lithium plated and stripped ddring each half-cycle. mAh1
n = Total number of cycles achieved
THERMAL STABILITY TESTS
The thermal stability tests were designed to evaluate the compatibility of
the cathode materials with the different electrolyte solutions. The samples
consisted of approximately 0.5 g of cathode material immersed in electrolyte
solution sealed in a glass ampul. Each sample also contained a strip of
lithium metal in contact with the solutions. The lithium strips were enveloped
in Celgard 2400 separator material to prevent direct contact with any particles
of cathode material that may have been dispersed in the solution.
Each cathode material/electrolyte solution combination was evaluated under
two storage conditions: Six months of ambient temperature (20-240 C) storage and
one month storage at 710 C followed by five months storage at ambient
temperature. After the storage intervals, the ampuls were opened in a glove
2-10
NSWC TR 86-108
box under an argon atmosphere. The electrolyte solutions were carefully
removed and analyzed by atomic absorption for either titanium or vanadium ionconcentration to define the solubility level of the cathode material. Thecathode materials were thoroughly rinsed with pure solvent to remove residualtraces of the electrolyte salts and then investigated by x-ray diffractionanalysis to determine if any gross changes in composition or structure hadoccurred.
Cycle Life Tests
Cycle life tests were conducted employing either half-cells (Figure 2-1) or2-plate wick laboratory cells (Figure 2-2) at both room temperature and-200 C. The separators used in laboratory cells consisted of two layers of0.001 inch thick Celgard 2400. Discharge and charge cutoff voltages werevaried in Li/TiS2 and Li/LixCoO 2 cells as a part of evaluations.
In the case of V205 , 3-platl wick laboratory cell was used with theutilized surface area of 6.4 cm . Table 2-2 describes the cathodes used in
these tests, while the test parameters are defined in Table 2-3. In thesetests, the depth of discharge was limited to a maximum of 0.75 F/mole (i.e., 75percent DOD). In this way, cells were discharged either 2.8 volts cutoff or
0.75 F/mole capacity limit, whichever was reached first.
2-11
NSWC TR 86-108
TABLE 2-2. CATHODE DESCRIPTION FOR V20 5 SCREENING STUDIES
TABLE 2-3. TEST PARAMETERS FOR V 205 CATHODE SCREENING STUDIES
Discharge Current Density: 1.0 mA/cm 2
Discharge Voltage Cutoff: 2.8 volts
Maximum Discharge Capacity Limit: 0.75 F/mole
Charge Current Density: 1.0 mA/cm 2
Charge Voltage Limit: 3.6 to 3.8 volts
Maximum Charge Capacity Limit: 0.75 F/mole
2-12
NSWC TR 86-108
CHAPTER 3
RESULTS AND DISCUSSION
CATHODE SELECTION STUDIES
Our overall objective in this work was to identify the cathode material/electrolyte solution combination offering the optimum capabilities towardmeeting the performance goals of the program. Four cathode materials have beenevaluated including TiS , V, V S and Li CoO . Our investigationsof ester-based electroljte goutign; focusedxpriiarily on two solvents: methylformate and methyl acetate. Iii addition, however, dimethyl sulfite (DMSI) wasincluded in some of the evaluations. Dimethyl sulfite has been successfullyused in primary lithium cells and has be n reported to behave as a reversibleliquid depolarizer at carbon electrodes. This latter quality could offerunique capabilities to rechargeable lithium applications, particularly withrespect to providing protection against overdischarge. and it was therefore ofinterest to include dimethyl sulfite in our investigations.
All solutions employed LiAsF 6 as the solute. LiAsF 6 was selected basedon its commercial availability in a high purity grade and its demonstratedperformance in rechargeable lithium applications. Methyl formate solutions.however, also containeg LiBF4 which is needed for solution stability atelevated temperatures.
To identify the best performing system, screening evaluations wereconducted as summarized in Table 3-1. Before use, each cathode material wascharacterized using the methodology outlined in Table 2-1. Stoichiometry,particle size, and purity can all have a significant impact on electrochemicalperformance and cyclability of insertion-type cathode materials. The purposeof the characterization tests, therefore, was to ensure, to the best possibledegree. that the cathode materials were of suitable quality before proceedingwith the screening evaluations.
Once the electrochemical system had been selected, work focused on cathodeprocessing, evaluating various manufacturing methods and then optimizing thecathode composition for performance. The effort under Phase I of the programculminated with a demonstration of the performance capabilities of the selectedsystem under extended cycle life conditions.
3-1
NSWC TR 86-108
TABLE 3-1. SUMMARY OF SCREENING EVALUATIONS CONDUCTED IN THIS PROGRAM
CATHODE MATERIAL
TEST SOLVENT NONE TiS 2 V701 V9S9 Li xCoO2
Cycle Life Tests Methyl Formate - X X - XThermal Stability Methyl Formate * X X - -
Lithium Cyclability Methyl Formate X . . ..
Cycle Life Tests Methyl Acetate - X X - XThermal Stability Methyl Acetate - X X - -
Specific tests were not conducted under this program. However, extensive
data were available from previous work conducted by Honeywell in unrelated
programs.
No work on V S due to its instability in ester-based electrolytes
(see Vanadium V) Sulfide section).
The following is a detailed discussion of the experimental results obtainedduring the selection, optimization, and demonstration of a rechargeable lithiumtechnology as conducted in this program.
CATHODE MATERIAL CHARACTERIZATION TESTS
Vanadium Pentoxide (V205)
Particle Size. The particle size distribution for V205 lots CML-V2-001and CML-V2-003 are shown in Figure 3-1. It was found that the particle sizedistribution of the manufactured V 05 correlated well with that of theNH4VO3 starting material, as illusirated in Figure 3-2.
The particle size distribution for Lot CML-V2-001. which employedas-received NH 4VO3 had major population between 76 p and 150 V. In an
effort to reduce the particle size of the manufactured V20 5 , Lot CML-V2-003employed NH4VO which had been ball milled. As can be seen in Figure 3-1.this caused th2 percentage of fine particles (i.e., < 45 V) to be significantlyincreased. However, a large portion of the population still fell in the76-150 14 range.
3-2
NSWC TR 86-108
PARTICLE SIZE ANALYSIS OF V205
Lot 001 LotO003
40Percent of Sample
35
30
25
20
15
10
0S<45 45-53 54-75 76-105 106-125 126-150 >150
Particle Size in Microns
FIGURE 3-1. PARTICLE SIZE DISTRIBUTIONS FOR V 20 5(LOTS CML-V2-O01 ANDl CML-V2-003)
3-3
NSWC TR 86-108
PARTICLE SIZE ANALYSISV206 vs. NH4VO3
Piwemt of lamp!.
35
30
35
20
is"VA
16 VA',- /10
* mpaaue ft" In Mms.
FIGURE 3-2. PARTICLE SIZE DISTRIBUTION FOR V2 05(LOT CML-V2-001) AND THE NH 4VO3 2USED IN ITS PREPARATION
3-4
NSWC TR 86-108
Morphology. Lot CML-V2-001 was used to characterize the morphology of the
manufactured V2o . Scanning electron micrographs for two sieve fractions
are shown in Figare 3-3. As can be seen. the morphology changes with the
particle size; the coarse particles (76-105V ) were observed to be ellipsoid in
shape, while the fine particles (, 45 1 ) tended to be much more irregular.
Stoichiometry. The stoichiometry of the manufactured V20 was
determined through a wet chemical analysis for total vanadium involving atitration of 0.lN ferrous ammonium sulfate using a barium diphenylaminesulfonate indicator. The results for the three V205 lots are as follows(theoretical = V02.5 0):
As can be seen, Lot CML-V2-001 was found to be in good agreement with the
theoretical value while the other two lots are indicated to be oxygen rich.
Crystal Structure. The x-ray diffraction data for the three lots ofV 2 0 are given in Tables 3-2 to 3-4. The results for all lots were foundto ie in good agreement with the Joint Committee on Powder DiffractionStandards (JCPDS).
Purity. The purity of the manufactured V20 was assessed through anEDAX analysis of Lot CML-V2-001. The only impuiity detected was a trace of
silicon, present at an estimated concentration of 0.6 weight percent.
Electrochemical Performance. Discharge results for the three lots of
V20 are shown in Figure 3-4. The cathodes used in these tests were made
by he cold pressing technique and contained 88 w/o V o + 10 w/o Dixon
Graphite + 2 w/o Teflon for Lot 001 and 90 w/o V20 + -i w/o ixon Graphite
for Lots 002 and 003. The cells were discharged ai 0.5 mA/cm to a 2.5 volt
cutoff using a 2M LiAsF6 + 0.4M LiBF4/MF electrolyte solution. Theseparator consisted of one layer of 0.010" thick HV-932 glass separator.
These tests employed 2-plate wick cells. Efficiencies of > 99 percent were
achieved in all cases, based on a theoretical capacity of 1.0 F/mole. The
average energy denrity was 459 Wh/kg based on active material.
Characterization Summary for V20. These characterization results show
that the manufactured V205 is a hi-hlv purified grade of material that can
yield essentially 100 percent of teoretical capacity at low discharge rates.
The particle size of V205 is indicated to be determined by the particle
size of the ammonium metavanadate from which it is manufactured and significant
morphological differences were observed between fine and coarse particles that
could influence the mechanical integrity of the finished electrodes and their
cycle life performance. Although two lots of material were indicated to be
oxygen rich, their discharge performance was excellent, demonstrating that the
indicated variation in stoichiometry does not significantly affect
electrochemical properties.
3-5
NSWC TR 86-108
0
-It
z4 In
zw0 o
> w
w'IU
3-6
NSWC TR 86-108
TABLE 3-2. X-RAY DIFFRACTION DATA FOR V2 05(LOT CML-V2-001)
d-spacing. A Relative Intensity, Percent
hkl JCPDS Standard Observed JCPDS Standard Observed
200 5.76 5.80 40 32
001 4.40 4.38 100 100
101 4.11 4.09 35 32
201 3.48 3.51 7 12
110 3.40 3.42 90 100
400 2.88 2.89 65 81
011 2.76 2.77 35 17
111 2.687 2.694 15 17
310 2.610 2.622 40 44- - 2.357 - 6
002 2.185 2.193 17 17
102 2.147 2.156 11 10
202 2.042 2.033 3 12
411 1.992 1.999 17 15
600 1.919 1.921 25 27
302 1.900 1.907 17 12
012 1.864 1.867 13 19
020 1.778 1.778 3 27
601 1.757 1.760 30 14
021 1.648 1.648 11 14
611 1.576 1.576 9 7
412 1.564 1.567 11 8
701 1.540 1.542 3 3
321 1.515 1.519 17 15
710 1.493 1.496 17 18
602 1.442 1.439 5 5
711 1.412 1.417 7 5
NOTES: 1. Analysis conducted under helium flow.
2. Observed relative intensity based on peak height rather
than on integrated peak area. o3. JCPDS Standard: a = 11.5A. b = 3.559A. c = 4.371k.
orthorhombic.
3-7
NSWC TR 86-108
TABLE 3-3. X-RAY DIFFRACTION DATA FOR V2 05(LOT CML-V2-002)
ad-spacing. A Relative Intensity. Percent
hkl JCPDS Standard Observed JCPDS Standard Observed
NOTES: I. Analysis conducted in air.2. Observed intensity based on peak height rather than
integrated peak area. 0 03. JCPDS Standard: a = 11.51A, b = 3.559A, c = 4.371A.
orthorhombic.
3-8
NSWC TR 86-108
TABLE 3-4. X-RAY DIFFRACTION DATA FOR V2 05(LOT CML-V2-003)
0
d-spacing, A Relative Intensity, Percent
hk JCPDS Standard Observed JCPDS Standard Observed
200 5.76 5.765 40 39
001 4.38 4.385 100 100
101 4.09 4.095 35 30
201 3.48 3.486 7 7
110 3.40 3.408 90 95
400 2.88 2.882 65 68
011 2.76 2.763 35 34
111 2.687 2.687 15 12
310 2.610 2.609 40 33
211 2.492 2.488 7 5
401 2.405 2.403 7 5
002 2.185 2.186 17 16
102 2.147 2.149 11 8
411 1.992 1.993 17 15
600 1.919 1.917 25 22
302 1.900 1.900 17 13
012 1.864 1.863 13 13
112 1.840 1.840 5 5
020 1.778 1.781 3 28
601 1.757 1.755 30 13- 1.701 - 4
021 1.648 1.649 11 13
121 1.632 1.634 7 5
611 1.576 1.575 9 8
412 1.564 1.563 11 9
701 1.540 1.538 3 5
321 1.515 1.514 17 14
710 1.493 1.492 17 17
602 1.442 1.443 5 5
711 1.412 1.413 7 6
022 1.380 1.380 5 5
303 1.363 1.362 5 5
NOTES: 1. Analysis conducted under helium flow.
2. Observed intensity based on peak height rather than
integrated peak area. o0 0
3. JCPDS Standard: a = 11.51A. b = 3.559A. c = 4.371A,
orthorhombic.
3-9
NSWC TR 86-108
* JVT
000'
H Ln
cC4
0~ u > >E
-c 0 NC1N H0
00
H ~ ~ -*D H~4-0-
* 0 0100 .r-4 HOu z
+ . + 0 +~
H~E- C-) Cl-4 0 .
-> o*~ ~ 0 . CLICJ -
00 c -) C,4 o E-
HOH0 0+a
V) Cl)C.
CD0 0 Z) -
0) 00 04 0
01 0 0 0 0n
E-4 -4
3-10
NSWC TR 86-108
Titanium Disulfide (TiS2)
Particle Size. The particle size distribution for the Degussa grade ofTiS is shown in Figure 3-5. Approximately 80 percent of the material fellat ihe extremes of the measured particle size range showing that the materialconsists primarily of a mixture of very large (>1501i) and very small (<45i)particles.
Morphology. SEM analyses were conducted on two segregated particle sizefractions; a coarse fraction (76-105V) and a fine fraction (<451i). Theresults, shown in Figure 3-6, show that both fractions are crystalline, butwith significantly different particle morphology. The coarse particles areellipsoid-shaped with some small, flake-like particles attached to the outersurface. The fine sieve fraction, however, consists almost entirely of thesmall flake-like particles.
Stoichiometry. The stoichiometry of Ti' was determined throughcombustion analysis at 9000 C. The results indicate that the Degussa grade ofTiS 2 is indeed stoichiometric, having a composition of Ti0.9992S2.
Crystal Structure. The x-ray powder diffraction pattern for TiS isgiven in Table 3-5. The d-spacings are in good agreement with the J6PDSstandard. The relative intensities in the observed pattern are also quitecompatible with the standard except that the two strongest peaks are reversed.This discrepancy is believed to be due to preferred orientations in the sample.
Purity. EDAX analysis detected no impurities in the Degussa grade of TiS 2 .
Electrochemical Performance. This test was conducted in a 2-plate wicklaboratory cell employing a 2M LiAsF 6 + 0.4M LiBF4/MF solution. Thecathode contained 80 w/o TiS2, 10 w/2 Dixon Graphite and 10 w/o Teflon andthe cell was discharged at 0.5 mA/cm to a 1.7V cutoff. The discharge curveis shown in Figure 3-7. The cell yielded an excellent capacity of 0.95 F/mole(theoretical = 1.0 F/mole) and the delivered energy density was 495 Wh/kg basedon active material.
Characterization Summary For TiS2. The analytical results demonstratethat the Degussa grade of TiS2 is of high pi:rity and stoichiometric. Theresults also show that this grade of material consists primarily of very smalland very large particles having significantly different morphologies.
Vanadium (V) Sulfide (V2S5)
Particle Size. No particle size analyses were conducted on themanufactured V2 S5.
Morphology. The morphology of the manufactured V2S5 was evaluatedthrough SEM analysis of Lot CML-VS-002. The material was found to consist oflarge crystals with small crystallites dispersed over their surface.Furthermore, as illustrated in the micrograph shown in Figure 3-8, themorphology of the V2S5 was found to be very similar to that of the(NH4)3VS4 starting material.
NOTES: 1. Analysis conducted under helium flow.2. Observed relative intensity based on peak height r .ther
than on integrated peak area. 03. JCPDS Standard: a = 3.4049X, c = 5.6912A. hexago-al.
3-14
NSWC TR 86-108
0v0
1-4Q
-'-
+ 0C4 WN
to C --
F-T
C'-CC,
C'C U- C ,
>0 E
- pE.
F- uF--
NZU~C~>
N~NCD
3-1-
NSWJC TR 86-108
.4
^ qp4 ~Ilk
4 1 z m-
4
%. .4"
3-16
NSWC TR 86-108
Stoichiometry. The stoichiometry of the manufactured V2S5 wasdetermined through combustion analysis in air at 6100 C to form V 205 . Theresults were as follows:
Lot CML-VS-002: V2S4. 3 3Lot CML-VS-003: V2S4.9 1
Lot CML-VS-002. prepared at 2600 C, is indicated to be sulfur deficient.The stoichiometry of Lot CML-VS-003, which was prepared at 2250 C. was muchcloser to the theoretical value indicating that lower conversion temperaturesare desirable in the manufacture of V2S 5 .
Crystal Structure. X-ray powder diffraction analysis of Lot CML-VS-002yielded no sharp peaks confirming that the manufactured V2S5 was amorphous.
Purity. No impurities were detected in Lot CML-VS-002 when analyzed byEDAX.
Electrochemical Discharge Performance. The cathodes used in these testscontained 80 w/o V S, 10 w/o Dixon Graphite and 10 w/o Teflon for Lot 2and 90 w/o V S ang 1; 1/o Dixon Graphite for Lot 3. The cells weredischarged ai 8.5 mA/cm to a 1.5V cutoff using a 2M LiAsF 6 + 0.4MLiBF4/MF solution.
The tests yielded delivered capacities of less than 2 F/mole with both lotsof V2 S (Figure 3-9). This output is significantly less than the 5 F/moletheoreiical capacity. The average energy density was 426 Wh/Kg based on activematerial. The solution and separator were both observed to be discolored inthe discharged cells.
Characterization Summary for V,)S. The results indicate that thetemperature used to convert (NHVS VS' to V S5 can affect the0 5stoichiometry of the product. At266 C, tie prepared material is indicatedto be sulfur deficient while at 2250C. the product had a stoichiometry inreasonable agreement with the theoretical value.
The cell performance achieved ith the two lots of V2S is very similarto that reported by Jacobsen et al 0 for cells cycled approximately 10 timesusing LiClO4 /dioxolane electrolyte solution. They postulate that theobserved capacity degradation may be due to the removal of edge sulfur atoms aslithium sulfide accompanied by polymerization to form amorphous VS2. Withester-based solutions, we believe that similar compositional changes areoccurring, but much more rapidly, being essentially complete by the end of thefirst discharge. The discoloration observed in the discharged cells furtherattests to compositional changes occurring in V2S5.
The compositional instability, apparently intrinsic to V 5 indicatesthat, in practice, a Li/V 2s system should be treated as a Li! S2system. Assuming a reversigle depth of discharge of 2.0 F/mole and an averagevoltage of 2.28 volts for VS P 10 the theoretical energy density thereforedecreases from 843 Wh/Kg to 443 Wh/Kg. thus offering no advantage over either
3-17
NSWC TR 86-108
08T
u~ 0 0>~ 0 0'
z0 0
Z0 Lf -401*
9 ~ ~ : U3 l
o0 0
E- E-4Cl + ! +CL u~ u. Cu
Cd) Cl))
F. u cz ~ 00 0'T
0
C.'I cn
0 -40C
cn 1-40 0 C-)0' C -3 .4 W
-4-
(~ -4
0 0 9
00 0 0'
00
0 0 0 0 00
0 0C0 0
3-18
NSWC TR 86-108
V205 or TiS,. This low energy density, combined with the difficulties inpro~essing F S due to its extreme sensitivity to moisture and oxygen.significantly diminishes the attractiveness of V 2S5* As a result, nofurther work was done with this cathode material.
Li x oO2
Particle Size. The particle size distribution for Lot CML-CO-002 is shownin Figure 3-10. The material was found to consist primarily of very large andvery small particles with the major fraction falling in the > 150 V range.
Morphology. SEM micrographs for the manufactured Li CoO LotCML-CO-002. are shown in Figure 3-11. The material was Touni to be highlycrystalline and the wide range of particle sizes are clearly shown.
Stoichinmetry. The stoichiometry of Li CoO was determined throughatomic absorption analysis for lithium and cobait. The results were as follows:
Lot CML-00-002: Li 85Co0 9702Lot CML-CO-003: Li1 .001 0002
Crystal Structure. The x-ray powder diffraction data for the two lots ofLi CoO 2 are given in Tables 3-6 and 3-7. The results for both lots are ingood agreement with the JCPDS standard.
Purity. EDAX analysis of LiCO Co Lot CML-CO-002. detected noimpurities in the manufactured material.
Electrochemical Performance. Li Co0 is prepared in the dischargedstate, thus necessitating a cell to te carged before its electrochemicalperformance can be evaluated. The following paragraphs describe tests thatwere conducted as part of the characterization of Lot CML-CO-002 material. Allcells employed aluminum current collectors in the cathode to prevent corrosionproblems at the high oxidizing potentials present with this system.
Figure 3-12 shows the initial charge and discharge performance of a Li/Li CoO 2-plate wick cell at 0.5 mA/cm employing a 2M LiAsF +0. M LiBF /MP solution. The cathode consisted of 80 w/o LoO 2. 10 w/oShawinigan Acetylene Black (SAB, 50 percent compressed) and 10 w/o Teflon, andwas prepared by the cold pressing technique. Charging was accomplished on a
test station with the upper voltage limit set at 5.0 volts. However, no upward
deflection in the voltage occurred to signal a fully charged state. As aresult, the cell potential never reached 5 volts so the charge was manually
terminated after a charge capacity of 1 F/mole. This represents approximately15 percent overcharge based on the measured stoichiometry of the startingcathode material. At the end of the charge period, the electrolyte solution
was observed to be dark amber in color, indicating that some decomposition ofthe solution had occurred. On the subsequent discharge, the cell deliveredonly 56 percent of the theoretical capacity, based on 1.0 F/mole.
3-19
NSWC TR 86-108
PARTICLE SIZE ANALYSIS OF LixCoO2Lot CML-CO-002
70 Per ent of Sample
60
50
40
30
20
10
•<4f- 46-68 8J4-70 76-10 06 -1=-IM M-100 >180
Pmidow a" ta Ekoa
FIGURE 3-10. PARTICLE SIZE DISTRIBUTION FOR Li xCoO 2
(LOT CML-CO-002)
3-20
NSWC TR 86-108
0
rQ
z
IF ,
3-2-
NSWC TR 86-108
TABLE 3-6. X-RAY DIFFRACTION DATA FOR Li coo(LOT ML-CO-002) x 2
d-spacing, A Relative Intensity, Percenthkl JCPDS Standard Observed JCPDS Standard Observed
NOTES: 1. Analysis conducted under helium flow.2. Observed relative intensity based on peak height rather
than on integrated peak area. 03. JCPDS Standard: a = 2.8166X. c =14.052A. Hexagonal
(rhonbohedral).
3-23
NSWC TR 86-108
00'
09*1
0
zfl00
0 +-
E n E 17
0 fn
0) 0 c 0 0
> E-4 0 ~0 CI >
~000
w 141
.4 0
-4 E- 09*
<N~~ 2 J74
0n .cD3
0. , wUw
.-~ .~ .0 0CD C
SlqO 0qI~a~ q0 E-
3-24~
NSWC TR 86-108
Since the low discharge capacity obtaiihed in the first test could have beendue to the observed electrolyte decomposition, additional tests were conductedto investigate the effects of solution composition on cell performance. Thethree solutions employed were:
2M LiAsF 6 + 0.4M LiBF 4/MF (baseline solution)
2M LiAsF6 /F
1.5M LiAsF 6 /MA
These tests were designed to determine (1) whether or not LiBF 4 was thecause of the observed solution decomposition and (2) if methyl acetate
solutions could offer better performance with Li CoO The cold pressedcathodes used in these tests consisted of 90 w/oxLi XoO2 and 10 w/oTeflon. No conductive diluent was employed in order to eliminate any possiblecarbon-catalyzed decomposition reactions. All Lesting was done atcharge/discharge rate of 0.5 mA/cm .
The initial charge/discharge curves for the three cells are shown inFigures 3-13 to 3-15. As with the first test, no upward deflection in voltagewas observed in any of the cells during charging to indicate a fully chargedcondition. Therefore, charging was terminated at a capacity of approximately1 F/mole. The following observations summarize the charging tests.
o In general, the charging curves consisted initially of a relativelyconstant voltage (below 4.0 volts) for a duration dependent on thesolvent medium. The cell employing the methyl acetate solutionreceived the greatest charge in this region.
o After charging to about 0.8 F/mole, the cells employing the methyl
formate solutions both exhibited irregularities in the voltage/timecurves indicating the possible onset of decomposition reactions due to
overcharge. The amber color of the solutions at the end of chargeconfirm that some solution decomposition had occurred. Also, since
decomposition was observed in both solutions, LiBF 4 is eliminated as
the source of instability.
o The cell employing the methyl acetate solution exhibited a smoothvoltage profile throughout the charging process and the solutionremained clear indicating that methyl acetate solutions may offersuperior capabilities for use with LixCoo 2 .
The discharge performance of .:he three cells is summarized in Table 3-8.Both cells employing methyl formate solutions yielded low capacities,
apparently due to the observed solution decomposition during charge. The cellincorporating the methyl acetate solution yielded a much better efficiency
along with an excellent energy density of 841 Wh/Kg (based on active componentsonly). The fact that this latter cell delivered a capacity very close to thelithium content of the starting cathode material suggests that the maximum
depth of discharge that can be achieved with Li CoO 2 may be determined by
the stoichiometry of the material when manufactured.
3-25
NSWC TR 86-108
00,0
08-0
u0
-4C
OVT N
zz0- -4
E -E- Q~ L-
C4 <0
C-44
0-4
04 n1
a4 -3 0 wx .I
(-4
CD CD 08
0C
SIO :IIN30 qq.0
Z F--- 3-26Z
NSWC TR 86-11)8
09,T
0 C00
00Li.. CU
CN
z,Iv -..- n oo
-- F 40 *
-- T Uj 0
C) 00080
-u 08
09*
zz 0 V
P.j u0
SYIOA < 713
3-27
NSWC TR 86-108
z 09'T
crri
cn 0T
:3 144 oz 0 -0
0
u-~ EnJ x -4
0 00 000' > (n0
> C)
0 ~ 09w V)Q
z E-4
0~~ 0
09
SI'IOA ~ C' >1INL~ V
3-28x
NSWC TR 86-108
TABLE 3-8. DISCHARGE RESULTS FOR Li/Li xCoo 2 CELLS
Delivered Average EnergyCapacity, Voltage. Density.
NOTES: t. The above performance values are based on a 3.OV cutoff.2. The cathode composition was 90 w/o Li CoO 2 + 10 w/o
Teflon binder. (Li CoO = Lot CML-CO-O02).3. The energy densityxis gased on active material only.
Characterization Summary of Li Co0 2 . The analytical results show thatthe manufactured Li CoO 2 is of higf purity. However, the reason for thedifference in stoiciiometry between the two lots of material is not clear.Evaluation of the electrochemical performance of the manufactured Li xCoO 2was complicated by apparent electrolyte solution decomposition duringcharging. Although improved performance was achieved using methyl acetate
solutions in place of methyl formate solutions, the absence of a voltagedeflection at the end of charge indicates that a solution electrolysis reactionmay still be accompanying the charging process. Therefore, the ability tocharge this high voltage cathode material remains a key concern, even with the
more electrochemically stable ester-based solutions. These initial resultsalso indicate that the depth of discharge achievable with Li CoO2 may be
determined by the stoichiometry of the starting cathode material.
YCLE LIFE TESTS
The objective of the cycle life tests was to define the reversibility ofeach cathode material in each of the electrolyte solutions. Due to the many
cells involved, the tests were limited to the minimum number of cycles neededto reach a conclusion on each system. Generally. 10 to 40 cycles were
sufficient for this purpose. Because at this point in the program, cathode
composition and manufacturing processes had not been optimized for the various
cathode materials.
3-29
NSWC TR 86-10E
V202 5
The cycle life performance for the V205 laboratory cells is shown in Figures3-16 to 3-18. At ambient temperature (20-24°C) the methyl formate solutions
gave the best performance where the capacity remained constant at the 0.75F/mole level. Cells employing methyl acetate and dimethyl sulfite solutionsshowed a sharp decline in capacity after 8 to 10 cycles. The loss in capacitywith the dimethyl sulfite solutions appears to be caused by a loss of chargingefficiency. Beginning with the fourth charge, a peak appeared in the chargingvoltage profile (Figure 3-19) and on the subsequent discharge the cell behavedas though it had only been partially charged, although the full charge capacityhad been delivered to the cell. These results indicate that electrolysis ofdimethyl sulfite may be occurring during the charging of V205 cells.
At -200 C, all of the solvents exhibited a dip in capacity in the first 4to 6 cycles. This dip is attributed to incomplete charging of the cells causedby the 3.6V limit. After the first few cycles, the charging cutoff voltage wasincreased to 3.8V, which greatly improved cell performance. Overall, the bestperformance was achieved with methyl formate solutions where the capacityleveled out at approximately 0.58 F/mole. Dimethyl sulfite also performedrelatively well at the low temperature.
Based on these results, methyl formate is clearly shown to be the superior,
solvent for use with. 2 05 " A typical discharge-charge curve with themethyl formate solution is shown in Figure 3-20. Furthermore, when used withmethyl formate solutions, V20 has been demonstrated to offer good cyclelife capabilities with no deg~adation in capacity observed in the limited cyclelife tests conducted.
3-30
NSWoC TR 86--1(,8
.99AABIENT
.60 - -20 0 C
< .46
S.29
CYCLE NT" ER
CHARGE/DISCHARGE RATE: 1.0 -,iA/c
DEPTH OF DISCHL,\RGE: 0.75 F-/ole
DISCHARGE VOLTAGE LIMIT: 2.8 Volts
CHARCP VOLTAGE LIMITS: AMBTENT TFMPFRATURE (20-240 C)
SEPARATOR: 2 layers of Celgard 2400 (0.001" thick each)
FIGURE 3-18. V205 SCREENING STUDIES: CYCLE LIFE PERFORMANCE
WITH 1M LiAsF6/DMSI SOLUTION
3-33
NSWC TR 86-108
w0. 00
00 (1)4l
P-4 q-:3
trn
E- 04-
+0 0 U..
C4
:3:>
0 -0
0. coP0 -
0'7.
-Z4 U4 E-
SflOA ~ C ctLN30 T3
3-34
NSWC TR 86-108
U 08*1 /
UlU
0
00
0+ w
ol 0 c>~ 6.
0 :00C -14
00. ..
CC
1u 1000
1-4H
EC
0) w
0 4fC -
H 00
H .. ... 0c
C C
3-35
NSWC TR 86-108
TiS 2
Methyl Formate Solutions. Li/TiS 2 laboratory cell tests employing a 2MLiAsF 6 + 0.4M LiBF4 /MF solution demonstrated a sharp decline in capacity inearly cycle life. This capacity degradation was observed regardless of cathode
process or cathode composition. Typical results are shown in Figure 3-21.
The charge/discharge voltage profiles observed with methyl formatesolutions were unusual and suggested a mechanism more complex than merelyinserting and removing lithium ions from the layered TiS 2 structure. During
discharge, no "knee" was observed at the end, even after delivered capacities
of up to 2 F/mole (Figure 3-22). During charge, distinct voltage plateaus were
observed above about 2.6V with similar plateaus appearing in the next discharge
(Figure 3-23). In addition, cathodes were observed to swell as much as 70
percent in methyl formate solutions as compared to approximately 5 percent in2-methyl THF solutions under similar conditions.
To help elucidate the cause for the poor cycle life performance in methyl
formate solutions, x-ray analyses were conducted on discharged and cycledcathodes. After the first discharge, a strong peak was observed having ad-spacing of 101 indicative of significant lattice expansion. After extendedcycling, cathodes exhibited only a few low intensity peaks indicating a major
degradation of the structure of TiS2 (Figure 3-24).
Based on these results, we believe that solvent cointercalation isoccurring with methyl formate solutions, resulting in the observed capacitydegradation. Therefore methyl formate solutions were deemed unsuitable for usewith TiS 2 in long cycle life applications.
Methyl Acetate Solutions. Initial testing with methyl acetate solutionsdemonstrated significantly improved performance over that achieved with methyl
formate. The voltage profiles were normal with distinct "knees" observed atthe end of discharge and smooth s-shaped profiles observed during charge(Figure 3-25). With 1.5M and 2.OM LiAsF6 /MA solutions, however, a gradualdecline in capacity occurred throughout cycle life (Figure 3-26). The rate ofthis capacity degradation was still too great to meet the cycle liferequirements of this program.
One approach evaluated to improve cycle life performance in methyl acetatesolutions was to employ higher solute concentrations. The lithium cyclabilitystudies discussed later in this chapter ("Lithium Cycling Efficiency
Measurements") showed that, with methyl acetate, lithium cycling efficienciesincreased with increasing solute concentration, suggesting that the salt may
Figure 3-27 shows cycle life performance for Li/TiS 2 half-cells employingLiAsF 6 concentrations over the range of 2M to 3M. Half-cells were employedin these studies to ensure that anode effects were not contributing to theobserved capacity degradation in the methyl acetate solutions. The cells werecycled between the limits of 2.8V and 1.7V based on cathode-to-referencevoltage readings. The results show that improved performance was indeedachieved with the 2.5M solution, although capacity decline was still observed.The 3M solution delivered lower capacity, but significantly, showed nodegradation in capacity over cycle life.
Because the 3M LiA"F6 /MA solutions are highly viscous, polarizationeffects are more pronounced. Therefore, additional tests were conducted withthe 3M solutions in which the voltage limits were extended to 3.0V and 1.4V.As can be seen in Figure 3-28. enhanced capacity was indeed achieved at ambienttemperature (20-240 C) with little degradation in capacity over cycle life.
At -200 C, however, the viscosity severely limits performance, even when employingthe wider voltage limits.
Extended cycling results for the 3M solution tested at ambient temperaturebetween the limits 2.8V and 1.7V are shown in Figure 3-29. Over 200 cycleshave been achieved demonstrating that long cycle life capabilities can berealized with TiS 2 in ester-based solutions. The high viscosity of the 3MLiAsF6 /MA solutions, however, severely limits low temperature performance andsimilar limitations would be expected for rate capabilities. Therefore, thesesolutions are not suited to meet the demanding performance requirements neededfor this program.
A second approach investigated to improve the cycie life performance ofTiS2 in methyl acetate solutions was to employ ether co-solvents. 9 The
co-solvents evaluated were 1,2-dimethoxyethane (DME), tetrahydrofuran (THF).and 2-methyl THF. These tests were also conducted in half-cells.
The cycle life results are shown in Figures 3-30 to 3-32. No beneficial
effects were achieved with DME. Both THF and 2-methyl THF, however, improvedcycle life performance demonstrating that ether co-solvents can indeed improvethe performance of TiS2 cathodes in methyl acetate solutions.
Li CoO 2
The initial performance evaluations of Li CoO 2 conducted as part of thematerial characterization studies demonstratel that decomposition of methylformate solutions occurs during charging with subsequent degradation indischarge performance. Methyl acetate solutions, however, appeared to be moreresistent to decomposition, thus offering encouragement for the continuedevaluation of this system. Those initial tests also indicated that the maximumcapacity obtainable with Li CoO 2 may be limited by the stoichiometry of thestarting material.
3-43
NSWC TR 86-108
E- C
00,00
-40-
uz
C::U
0 U-
C+NC
.*-.
0- u
E-4 Ln-
CN -jC E-4=Z
cn ru Lr
3-4-4
NSWC TR 86-108
N(
C)~ w) C):
E-
000
-4
00
V rV)
C)C
000
* 00*01wV
3~~~IOW/d~~ :z~ycV 3~VJ
3-4-4
NSWC TR 86-108
000
Q0
E-3
0 z4
z LI.]
03 u
E _ >0~ 00
zz;4 w 0
EU C 0000or.u -o
CZ Cn E-4 0E-
-3-4
NSWC TR 86-108
ee.0s
zz
-4z EA 0
CN CE
Z
.. II
u OC / e4eCw I=E- /r
+ E-
+ U)
u u -92 /
V) W
CZ
>~ C)
0 ~0
C
0010
co CD w -
310wA :iovavo 3D91VHoSIc
3-47
NSWC TR 86-108
Ic z
w > C4
z FC
LN N
00z~ crIE V1L
00C
I-.
C,;
F-~L w, i/F
/c 8 0L
A 000 ~
CC) CD V
]IOI I.,.dv loE-SI
3-4
NSWC TR 86-108
00 0 ,
00'OE
Ic I
= z
LL1
000
-4.-49
NSWC TR 86-108
To better define the performance capabilities of Li CoO , additionalcycle tests were carried out focusing on methyl acetate electrolyte solutions.Dimethyl sulfite solutions were not considered for use with Li1 CoO2 becauseof their lower electrocheical stability at high oxidizing potentials.
The majority of the cycle life tests were conducted in half-cells so as toeliminate any anode polarization effects in the methyl acetate solutions. Thecathodes were manufactured by a cold pressing technique and contained 10 weightpercent Shawinigan Acetylene Black (50 percent compressed) and 4 weight percentisotactic-polypropylene binder.
Because Li CoO 2 cells do not exhibit a voltage deflection at end ofcharge, it is lifficult to establish an upper voltage limit for this system.If the voltage is allowed to go too high, solution decomposition isinevitable. If the voltage limit is set too low, however, the chargingcapacity will be limited.
Therefore, various charging cutoff voltages were evaluated in the ranIe of4.5V to 4.85V. In all tests, a low charging current density of 0.5 mA/cmwas employed so as to minimize polarization effects and maintain the chargingvoltages as low as possible.
Figuri 3-33 shows the results for half-cells tested at a d'scharge rate of1.0 mA/cm . As can be seen, all cells delivered an initial capacity of lessthan 0.6 F/mole and no additional capacity was obtained from the cell employingthe higher stoichiometry material (i.e., Lot CML-CO-003, LiI 0 ?Co 0002). Allcells exhibited a significant loss in capacity after z few ccres Accompaniedby discoloration of the electrolyte solution.
Figure 3-34 shows the results for a half-cell and a 2-plate laboratory .ickcell tested at a discharge rate of 5 mA/cm . 2The initial capacities wereapproximately the same as obtained at 1 mA/cm indicating that Li CoO 2does indeed possess good rate capabilities. However, a rapid decline incapacity was again observed in these cells accompanied by discoloration f theelectrolyte solution. The fact that the laboratory cell delivered fou- cyclesbefore any loss in capacity was observed demonstrates that LixCoo2 docsoperate reversibly in the absence of solution decomposition.
Although only limited testing has been completed with Li CoO 2, :heresults demonstrate that electrolyte decomposition remains alkey problem thatmust be resolved before this system can be practical. Although someimprovement is indicated when the charging voltage is limited to 4.5V to 4.6V,the cycle life performance obtained is still disappointing. Th-refore. in itspresent stage of development. LixCoO 2 could not be considered viablecandidate for this program.
3-50
NSWC TR 86-108
~ 00,91
<0 w-
-- s
0 0F-
00*01c CC)
00 E 00J .. ZC N00,8 F
U C4 IF-I11
~ ) 0-
F- ~ < / F-CCCA C) w -
<-~~E-- E- /
/0 -
3-51
NSWC TR 86-108
S00*21
0000
CD 0 C 00,8
30 12
00 0
z 00 0/9
/ ~ -~008
>~-0 0c-2
00*0
3-522
NSWC TR 96-108
TFERMAL STABILITY TESTS
Electrolyte Solutions
Although no separate tests were conducted in this program regarding thethermal stability of the individual electrolyte solutiozis themselves, anextensive data base has been developed at Honeywell on the elevated temperaturestability of LiAsF 6/MF and LiAsF /DMSI solutions that can supplement ourother screening test results invglving selection of the electrolyte solution.Our work in this area has shown that both methyl formate and dimethyl sulfitesolutions require contact with lithium metal for stability gt elevatedtemperatures. Methyl formate solutinns also require LiBF 4. When thesecondi:ions are met, excellent stability is achieved, as summarized in Tables3-9 and 3-10.
TABLE 3-9. AMPUL STORAGE TEST RESULTS FOR 2M LiAsF +0.4M LiBF 4 /MF SOLUTIONS (HISTORICAL DATk)
Lithium ElectrolyteQuantity, Water Level, Storage Temperature, °C
NOTES: 1. The above values represent days required for 50 percept ofthe ampuls in a test group to fracture.
2. Each test group contained 50 ampuls.
3. Ampul fracture is caused by gas-producing decompositionreactions of the electrolyte solution. Therefore, the timerequired for an ampul to fracture gives a direct indicationof the rate that solution decomposition is occurring. Thepoint at which 50 percent of the ampuls have fractured 4sused to define the average rate of decom;)sition for a
particular test group.
3-53
NSWC TR 86-108
TABLE 3-10. AMPUL STORAGE TEST RESULTS FOR 1.0M LiAsF 6 /DMSI
SOLUTIONS (HISTORICAL DATA)
LithiumQuantity, Storage Temperature, °C
mg/ml 64 74 84
0.5 1559 532 179
2.1 2790 846 262
6.2 >3024 1475 323
NOTES: 1. The above values represent days required for 50 percent of the
ampuls in a test group to fracture.
2. Each test group contained 50 ampuls.
3. Ampul fracture is caused by gas-produuing decomposition reactions
of the electrolyte solution. Therefore, the time required for anampul to fracture gives a direct indication of the rate that
solution decomposition is occurring. The point at which 50 percent
of the ampuls have fractured is used to define the average rate of
decomposition tor a particular test group.
These historical results, therefore, demonstrate that methyl formate anduimethyl sulfite solutions offer suitable thermal stability for ourrechargeable applications. Unfortunately, however, similar data are not
available fcr methyl acetate solutions.
Cathode Materials
V 05 . Thermal stability tests with V2C5 were conducted in methylclla methyl acetate solutions only. The physical appearance of the
samples after storage are summarized in Table 3-11. At ambient temperature no
changes were observed, while at 71 C, significant changes were obseived in
both solutions.
The results of the solution analyses are shown in Table 3-12. At ambient
temperature, no solubility is indicated in either solution. At 710C.
K was found to be insoluble in the 2.OM LiAsF 6 + 0.4M LiBF 4 /MF-clution. With the 2.0M L iAsF.,,MA solutions stored at elevated temperatures.
the lithium metal was completely consumed and the ampuls were observed to
.:ntain significant pressure when they were opened. The V O had changed
color from orange to hlack and the solution contained approximately 800 ppm
AA rcs 'tlLiol! in these "samplcs i ; 2.5 ppm for Vanadiumi.
3-55
NSWC TR 86-108
X-ray diffraction data for the recovered cathcee material samples are givenin Tables 3-13 to 3-16. All results are in good agreement with the JCPDSstandard except for the sample stored in 2.OM LiAsF 6 /MA at 710C. The datafor this laHer sample indicate that the V205 had been largely converted to= -Li/V 2 05
Based on these results, it is concluded that V205 is chemically
covpatible with the 2.OM LiAsF 6 + 0.4 LiBF /MF solution, but not with the2.014 LiAsF 6'MA solution. The6 frato o ihae V 1 o in the lattersolution during elevated temperature storage was quite ~u prising and indicatesthat some type of reducing agent may have been generated in the solution.
TiS 2
Thermal stability tests with TiS 2 were conducted in methyl formate andmethyl acetate solutions only. The physical appearance of the samples after
storage are summarized in Table 3-17. No changes were observed in the samplesmaintained at room temperature. At 710 C, however, significant changes wereobserved, particularly with the 2M LiAsF6/MA solutions.
The results of the solution analysis for titanium concentration are shownin Table 3-18. No solubility is indicated in either solution at roomtemperature nor in the methyl formate solution stored at 710 C. The methylacetate solutions at elevated temperatures contained approximately 170 ppmtitanium indicating some solubility of TiS However, these solutionscontained a fine suspension which could noi be removed even by filtration.Therefore, it is not clear how much of the detected titanium came fromdissolved material and how much was due to the suspended solids.
The x-ray results of stored TiS, were complex. Only the sample stored in2.OM LiAsF /MA at ambient temperature gave a normal TiS2 pattern. Allother samp~es exhibited numerous additional peaks indicating the presence of
other phases in addition to TiS2, as shown in Tables 3-19 to 3-21.
Based on these results, TiS 2 has been found to be insoluble in methylformate solutions qnd, possibly, slightly soluble in methyl acetate solutions.The chemical compatibility of TiS in these solutions, however, is notconclusive due to the detection oi additional phases in the stored samples.
LITHIUM CYCLING EFFICIENCY MEASUREMENTS
\lI lithium cvc ing efficiencie were determined at strip and plate curren
,loai ties of I mA/cm at ambient temperature (20-24 C). Otur objective in chis
t sk v-, to identify an ester-based electrolvte solution that could offer a lithium
v,-ino -fficiencv of ) 91.6 percent, the value reported for a LiAqF6 1 y-ethyl THF
so lt ion uinder test conditions ,I milar to those that we were employing. At the
, ,- ct, * we reaiz7ed that it was unlike lv that any of the pure solutions could achieve
ci is level of performance. Therofore, our qpproach was to etablish the baseline
Ctf -i ' e ahievable with ea;ch s-olvent and then investigate the use of additives
ho , r -o- lv,n ts as a meonq s )f fur t her enhancing' performance to the desired
NSWC TR 86-108
TABLE 3-13. X-RAY DIFFRACTION DATA FOR V2 05 STORED IN 2M LiAsF 6
+ 0.4M LiBF4 /MF AT AMBIENT TEMPERATURE (20-24°C)
d-spacing, A Relative Intensity. Percent
hkl JCPDS Standard Observed JCPDS Standard Observed
200 5.76 5.776 40 54
001 4.38 4.370 100 100
101 4.09 4.084 35 47
110 3,40 3.404 90 48
400 2.88 2.881 65 54
011 2.76 2.762 35 30
1il 2.687 2.684 15 7
310 2.610 2.611 40 27
002 2.185 2.183 17 13
102 2.147 2.142 11 8
411 1.992 1.993 17 15600 1.919 1.916 17 20
302 1.900 1.901 17 14
012 1.864 1.861 13 14
020 1.778 1.781 3 24
601 1.757 1.755 30 8
021 1.648 1.648 11 12
121 1.632 1.630 7 6
412 1.564 1.564 11 8
701 1.5396 1.5372 3 7
321 1.5149 1.5149 17 6
710 1.4925 1.4909 17 7
711 1.4090 1.4123 7 6
NOTES: 1. Analysis conducted under helium flow.2. Observed relative intensity based on peak height rather than on
integrated density. o o 0
3. JCPDS Standard (V205 ): a = 11.51A, b = 3.559A. c = 4.371A.orthorhombic.
3-57
NSWC TR 86-108
TABLE 3-14. X-RAY DIFFRACTION DATA FOR V2 0 STORED IN 2M LiAsF6 /MAAT AMBIENT TEMPERATURE (20-240
0d-spacing. A Relative Intensity, Percent
hkl JCPDS Standard Observed JCPDS Standard Observed
2. Observed intensity based on peak height rather than on integratedpeak area. 0 0
3. JCPDS Standard (TiS2 ): a = 3.4049A. c 5.6912A, hexagonal.
3-65
NSWC TR 86-108
levels. We also felt that solute concentration could have a significant impacton lithium cycling efficiency and therefore included this variable in ourevaluations.
Methyl Formate Solutions
Baseline Evaluations. The objectives of these initial tests were todetermine the cycling efficiencies of 2M LiAsF /MF solutions and to establishthe effects of the following parameters on lithium cyclability:
o LiBF4o Cell Configurationo Solution Deaeration with Argono Substrate Material
The results are summarized in Table 3-22. It was found that neithersubstrate material nor solution deaeration had a significant effect on lithiumcycling efficiency. Flooded cells, however, tended to give lower results than2-plate wick cells. LiBF 4 was indicated to be somewhat detrimental, lowerin&the efficiency from approximately 77 percent to about 70 percent when added ata concentration of 0.4M. Typical voltage profiles for the solutions with andwithout LiBF 4 present are shown in Figures 3-35 and 3-36.
The precipitates formed during the cycling tests were collected andanalyzed by infrared spectroscopy using KBr pellets. The resulting spectra areshown in Figures 3-37 and 3-38. Th? strong and broad peaks around 3400 cm-1
and the sharp peak at 1620-1640 cm- are attributed to moisture, apparentlyintroduced during sample preparation and handling. Characteristic peaks due toAsF6 were present in both spectra at 710-720 cm- and around 400 cm-1 .Also., in the spectrum of the material collected from the 2M LiAsF6 + 0.4M LiBF4 /MF solution, peaks were observed at 1080 and 520 cm . characteristic of theBF 4 anion. Table 3-23 bummarizes the bands observed in each sample.These analyses indicate that both the solvent and solute are involved in theinteraction with lithium during extensive cycling.
Based on these results, the wick cell was selected as the preferred cellconfiguration with methyl formate solutions. It was also evident thatsignificant improvements were required in the lithium cycling efficiency ifmethyl formate solutions were to be practical in rechargeable applications.
Effects of Solute Concentration. Table 3-24 summarizes the lithium cyclingefficiencies achieved in methyl formate solutions as a function of LiAsFconcentration over the range of 0.5M to 2M. The results, shown graphica~ly inFigure 3-39, demonstrate that the lithium cycling efficiency increasessignificantly as the solute concentration is lowered.
Although 2M LiAsF 6 /KF solutions would be preferred in order to achievemaximum performance capabilities, methyl formate solutions are so conductivethat even dilute solutions can offer reasonable conductivities. For example, a0.5M LiAsF6 /F solution has a conductivity of 16.3 mMHO/cm at ambient
NSWC TR 86-108
TABLE 3-22. LITHIUM CYCLING RESULTS IN MFTHYL FORMATESOLUTIONS --BASELINE EVALUATIONS
Represents the average of two experiments. Individual results are given in
Table 3-22.
3-72
NSWC TR 86-108
I z
0 +
ocC1 z 4-
E- 0 *
H z
0-4
o z U.+/ .
I 4..2 .
. I • uI ii i-- iiz
3-73
NSWC TR 86-108
temperature. Therefore, lower solution concentrations could indeed be a viablemeans of improving the cyclability of lithium in methyl formate solutionswithout severely degrading performance.
Effects of Electrolyte Additives.9 In an effort to improve the lithiumc):ling efficiency of concentrated methyl formate solutions, additivesconsisting of 2-methoxy-ethanol, 2-methyl-furan, and carbon dioxide (CO2)were evaluated. 2-methoxy-ethanol and 2-methyl-furan were selected based ontheir reportedi b ity to enhance the lithium cycling efficienzy in 2-methylTHF solutions. 1
C02 was evaluated as a precursor to reduce the parasitic loss of platedlithium Through the formation of an ionically conductive, protective film. Theinvestigations included using CO 2 both as an electrolyte additive and as amethod to pretreat the lithium anode prior to exposure to the electiolytesolution.
The results for 2-methoxy-ethanol and 2-methyl luran are given inTable 3-25. As can be aeen, these additives were both found to be detrimental.yielding significantly lower cycling efficiencies than the undoped solution.
Table 3-26 summarizes the results achieved with CO Pretreatment of theworking lithium electrodes was carried out by storing he electrodes in asealed vessel with 20 psig C02 pressure for 3, 24. or 120 hours. Thistechnique was found to significantly improve the lithium cycling efficiency inthe more concentrated solutions. Figure 3-40 illustrates this effect bycomparing the results for solutions with and without pretreatment of theanode. Using this approach, efficiencies of 84 to 87 percent were achievedwith a 1M LiAsF 6 + 0.21A, LiBF 4/MF solution.
In the tests employing CO2 as an electrolyte additive, solutions werefirst saturated with CO2 and then, following activation, the cells werepressurized with CO2 gas. As can be seen in Table 3-26. excellent resultswere achieved by this method. The lithium cycling efficiency of 93 percentobtained with a 2M LiAsF + 0.4M LiBF4/MF solution exceeds the reportedvalue for LiAsF 6/2-methv, THF solutions under similar experimentalconditions. 002 was also observed to decrease the anode polarization levels.as illustrated in Figures 3-41 and 3-42.
Methyl Acetate Solutions
BAeline Eval-ations. Table 3-27 summarizes the results for the baselinetests conducted Vith the 2M LiAsF /MA solutions. These results indicate alithium cycling efficiency of approximately 70 percent for this solution withno significant effects noted for solution deaeration, substrate material, orcell configuration.
The infrared spectrum for the solid product formed during the cycle testsis shown in Figure 3-43. while the observed bands are summarized inTable 3-28. These results indicate that both the solute and solvent axeinvolved in the reactions leading to the formation of the solid product.
3-7
NSWC TR 86-108
TABLE 3-25. EFFECT OF ORGANIC ADDITIVES ON LITHIUMCYCLING EFFICIENCY IN METHYL FORMATE
StartingAdditive WorkingConcen- Electrode Lithium
trate, Capacity. No. of EfficiencyElectrolyte Solution Additive w/o mAh Cycles Percent
NOTES: 1. All tests were conducted in wick cells employing nickel substratesfor the working electrodes.
2. All solutions were deaerated with argon prior to use.
3-85
NSWC TR 86-108
The results for CO pretreatment of the lithium surface are given inTable 3-32. As before, the pretreatment process consisted of storing thelithium working electrodes in a sealed vessel with 20 psig G02 pressure for3, 24, and 120 hours. In the methyl acetate solutions, however, pretreatmentof the lithium surface with CO2 was found to have no beneficial effect. Nofurther improvements were pursued due to the incompatibility of methyl acetatesolutions and V205 cathode material at elevated temperatures.
Dimethyl Sulfite Solutions
Baseline Evaluations. Initial testing was conducted using iMLiAsF6 /DMSI solutions. Half-cell tests yielded a lithium cycling efficiencyof 81 percent for this solution. However, satisfactory results could not beachieved in wick cells due to severe polarization and noisy voltage/timeprofiles. This could be explained by cell configuration. In the laboratorycell. formation of pale grey precipitates was observed on the Lithium surfaceof the working electrodes. In the half-cell, loose, solid products fo:med atthe working electrode can easily fall to the bottom of the cell tube thusminimizing passivation effects.
Figure 3-45 shows the infrared spectrum for the solid productcollected after the cycle tests while Table 3-33 summarizes the observedbands. As with the other solutions, both solvent and solute are indicated tobe involved in the formation of the solid product.
TABLE 3-32. EFFECT OF ELECIRODE PRETREATMENT WITH C02 ONLITHIUM CYCLING EFFICIENCY IN METHYL ACETATE
StartingStorage Time Workingfor Working Electrode LithiumElectrode in CO2 Capacity, No. of Lfficiency,Atmosphere, h mAh Cycles Percent
Effects of Solute Concentration. Table 3-34 summarizes the effects ofLiAsF 6 concentration on the lithium cycling efficiency in DMSI solutions overthe range of 0.5M to 2M. These tests were all conducted using half-cells. Theresults, which are graphed in Figure 3-46, indicate that the maximum lithiumcycling efficiency occurs at a LiAsF 6 concentration near IM.
SCREENING STUDIES SUMMARY
The objective of the screening studies was to identify the cathodematerial/electrolyte solution combination offering the optimum performance
capabilities for a rechargeable lithium cell technology. Based on the resultsobtained, the following system was selected.
o Cathode Material: V 0o Electrolyte Solution: 2A LiAsF 6 + 0.4M LiBF4/MF doped with CO2
This system offers good cycle life performance with little capacitydegradation combined with good thermal stability and high lithium cyclingefficiencies. In addition, the high conductivity of the methyl formatesolution offers the capabilities needed to achieve enhanced rate capabilitiesand low temperature performance over existing rechargeable lithiumtechnologies.
Results for the other candidate materials evaluated are summarized below.
LiAsF6/Methyl Acetate Solutions
LiAsF 6/methyl acetate solutions offer good conductivities and weretherefore very attractive for consideration in rechargeable cell applications.
3-88
NSWC TR 86-108
TABLE 3-34. EFFECT OF LiAsF CONCENTRATION ON LITHIUMCYCLING EFFICIECY IN DIMETHYL SULFITE(HALF-CELL. NO SOLUTION PRETREATMENT)
StartingWorking
Electrode LithiumElectrolyte Substrate Capacity, No. of EfficiencySolution Material mAh Cycles Percent
2M LiAsF6iDMSI Ni 54.1 19 76IM LiAsF /DMSI 304 SS 45.0 19 810.5M LiAsF6 /DMSI Ni 57.8 19 74
The major deficiency identified with these solutions was their low lithium
cycling efficiencies. The maximum efficiency achieved was only 73 percent andno improvements were realized with any of the additives or co-solventsevaluated. In addition, methyl acetate solutions were found to be unstablewith V 0 at elevated temperatures. Methyl acetate solutions, therefore.were fgu~d to be clearly inferior to the methyl formate solutions.
LiAsF 6 /Dimethyl Sulfite Solutions
Because of the limited testing done, the capabilities of dimethyl sulfite
for rechargeable applications have not been fully defined. However, theresults do indicate that dimethyl sulfite is not suitable for use with high
voltage systems as evidenced by the apparent electrolysis reactions observed
during charging of V205 cells.
TiS2
Poor cycle life performance was obtained with TiS2 using methyl formatesolutions, apparently due to solvent cointercalation. With methyl acetatesolutions, much improved performance was achieved. With these solutions, ithas been found that capacity degradation decreases as the solute concentrationincreases. Over 200 cycles have been achieved using a 3M LiAsF6/MAsolution. These concentrated solutions, however, are highly viscous whichseverely limits their rate capabilities and low temperature performance. Inaddition, the low lithium cycling efficiencies offered by methyl acetatesolutions are not sufficient for long cycle life applications. Therefore, theTiS 2/methyl acetate system does not offer the performance capabilities neededfor this program.
3-89
NSWC TR 86-108
Os . z
z
Z E-
0 . .-4
u4 U24
0 (n u
cz,
z Zw0
od
E- cn-
z
14 -AE-
ca0
3-904
NSWC TR 86-108
Li xCoO 2
The major problem encountered with this cathode material was electrolytedecomposition at the high potentials required during charging. Methyl acetatesolutions yielded better performance than methyl formate solutions butsignificant capacity degradation was still observed in early cycle life. Inaddition, the maximum discharge efficiencies typically achieved were only about
60 percent.
In spite of the disappointing results obtained in these preliminaryevaluations, we still feel that Li CoO is an attractive high energydensity cathode material. In the lbseice of electrolyte decomposition.Li CoO was indicated to operate reversibly and, even at an efficiency ofonly 66 percent. could yield an active material energy density of about670 Wh/kg. Thus. with proper development. Li CoO could indeed be a viablerechargeable technology. However, we felt thit tie magnitude of effortrequired to fully develop this system was beyond the scope of this program and.
therefore, were forced to discontinue work with Lix CoO 2 .
Our preliminary investigations showed V2S to be unstable with
ester-based solutions. Although detailed analyses were not conducted, webelieve that V S, decomposes to VS 2 in these solutions. Because of thisapparent inherin instability. V2 S5 was dropped from consideration early in
the program.
V2 05 CATHODE PROCESSING STUDIES
Introduction
Cathode processing is a critical part in the development of a practicalrechargeable lithium technology. To achieve efficient operation and long cyclelife capabilities, the physical properties of manufactured cathodes areextremely important. Cathodes must be porous to allow for adequate distri-bution of the electrolyte solution within the electrode structure. At the sametime, however, cathodes must be rugged and flexible so that they can be easilyhandled during cell manufacture and so that they do not deteriorate duringextensive cycling.
3-91
NSWC TR 86-108
To achieve ruggedness and flexibility, binders are essential. However,binders must be identified that can provide the desired physical propertieswithout degrading performance. Conductive diluents are also necessary toachieve good rate capabilities and low temperature operability. Conductivediluents also improve efficiency during extended cycling by helping to maintain
good particle-to-particle contact throughout the volume of the cathode.
Cathode processing methods must be developed that are tailored to thespecific properties of each cathode material and binder type. To ensure arapid development cycle, it is imperative that these processing methods be
production-oriented, suitable for scale-up with respect to both electrode sizeand electrode quantity.
Once V2 0 had been selected in the screening studies, effort focused ondeveloping stitable cathode manufacturing processes for this material. Theobjective of this work was to identify a production-oriented manufacturingtechnique that could produce rugged, flexible cathodes that operate efficientlyover a wide range of temperatures and discharge rates without significantdegradation in capacity over cycle life.
Process Evaluations
Evaluation of manufactured cathodes considered both mechanical properties
and electrochemical performance. Table 3-35 summarizes the cathodecompositions employed along with the mechanical properties of the finishedelectrodes. In most processes, Shawinigan Acetylene Black (50 percentcompressed) was employed as the conductive diluent. For the roll milledelectrodes, however, Vulcan XC-72R carbon was used because our experience inmanufacturing carbon electrodes has shown this carbon to be the easiest to rollmill. The ro,: milled electrodes also employed a different grade of V205than the cathodes made by other processes. This V^0 5 was obtained from ourG2666 cell production line and is manufactured in aermany by Metallurg. Theroll milling process requires a relatively large amount of material to make asingle cathode pad. Therefore, for these initial investigations, it wasdecided to employ the Metallurg material and conserve the V 0 that hadbeen specially synthesized and characterized in-house for lit~r studies. All
other processes. however, employed V205 that had been prepared in-house(Lot CML-V2-0O1).
From a mechanical standpoint, the cold pressed, Teflonated cathodes were
the least desirable, offering little in the way of ruggedness or flexibility.
Sintering. however, was found to greatly improve the properties of the cold
pressed electrodes. At the other extreme, roll milled cathodes were found to
be extremely rugged and totally flexible.
Performance evaluation of the manufactured electrodes was conducted through
limited crcle life tests in 2-plate wick cells at a charge/discharge rate of
1.0 mA/cm . A 2M LiAsF 6 + 0.4M LiBF 4/MF solution was employed in alltests.
3-92
L~..,,m m m m ~ m~mm m m ~ m
NSWC TR 86-108
TABLE 3-35. CATHODE DESCRIPTIONS FOR V2 05 CATHODE PROCESSING STUDIES
Initial
Discharge Cycle
Cathode Process Ruggedness Flexibility Performance Performance
Cold Pressing with
Teflon Binder Poor Poor Excellent Excellent
Cold Pressing and
Sintering with
Teflon Binder Good Good Good Good
Cold Pressing with
Isotactic Poly-
propylene Binder Fair Fair Good Good
Pasting with EPDM Binder Excellent Good Poor Poor
Roll Milling with
Teflon Binder Excellent Excellent Good Fair
NOTE: Performance evaluations based on limited cycle-life tests at a
charge/discharge rate of 1.0 mA/cm
The cycle life results are shown in Figure 3-47. It was found that the
cold pressed, Teflonated electrodes yielded the best performance. Their
average efficiency over 10 cycles was 87.3 percent (based on a theoretical
capacity of 1 F/mole) and little degradation of capacity was observed over
cycle life. The sintered electrodes also performed relatively well as did thepressed electrodes employing isotactic-polypropylene binder. The pasted
electrodes performed poorly while the roll milled electrodes yielded
intermediate performance.
The results in Figure 3-47 show that all processes evaluated exhibited
little degradation in capacity after the second cycle. In general, therefore.
the primary effect of cathode processing on performance of V O5 cathodes
appears to be in the magnitude of capacity loss between the first and second
discharge. Stated in another way, these results indicate that capacity decline
in early cycle life is an effect of the mechanical properties of the cathode
and not due to degradation of the cathode material. This is very significant
in that it suggests that this loss in capacity in the first few cycles can be
controlled, and possibly eliminated through optimization of the cathode
manufacturing process.
3-93
NSWC TR 86-108
zw
00 ol
00 9
U -
I cu CD U- E
z 0 -
o 0 ~ 0In. CL C000
I' ~ l UZC -Icn 0
I -94 <
NSWC TR 86-108
Summary
Based on the results of the evaluations of the five cathode processingmethods, it is concluded that the following two processes offer the bestcombination of mechanical and performance properties:
o Roll Milled Electrodeso Cold Pressed and Sintered Teflonated Electrodes
Roll milled cathodes are, by far. the best from a mechanical standpoint andare also suited to scale-up. Although their initial performance was somewhatdisappointing, the results suggested that through proper optimization, theseperformance deficiencies could be overcome. Therefore, roll milling wasselected as the process of choice. However. in the event that the performanceof roll milled cathodes could not be improved to the necessary level. sinteredelectrodes would then be employed.
V205 ROLL MILLED CATHODE OPTIMIZATION STUDIES
Introduction
To improve the performance of roll milled V205 cathodes, investigationswere conducted in an effort to identify and optimize the key parameter(s) thatcontrol the performance of these electrodes. The parameters evaluated includedcarbon type. carbon content, binder content, and V205 particle size.
Because of the relatively large quantities of V205 needed tomanufacture the numerous lots of roll milled cathodes, these studies were alsocarried out using the Metallurg grade of V2 05 from our G2666 cellproduction line.
These initial optimization studies were conducted in 2-plate wick cellsemploying a charge/discharge rate of 1.0 mA/cm . The electrolyte solutionwas 2.OM LiAsF 6 + 0.4M LiBF 4 /MF and all testing was done at ambienttemperature.
Evaluation Tests
Table 3-36 summarizes the physical properties of the manufactured cathodesalong with their cycle life performance results.
Effects of Carbon Type. In the initial investigations of roll milledcathodes, Vulcan XC-72R carbon was used as the conductive diluent because ourexperience had shown it to be well suited to the roll milling process.However, this carbon had not been used previously with V .o so its effecton performance was not known. Therefore, additional catiodes were manufactured
and tested with Shawinigan Acetylene Black (SAB) as the conductive diluent. Both50 and 100 percent compressed grades were evaluated.
3-95
NSWC TR 86-108
o a
C-4.
o 1.
C14 - c
>4 N4
o 0 0
0- 14011 14p
E--
- - 0 0u
1 .4 - . C
0 * '
0 0
0 4 11 *1C
-J 0 0 4 1
0 Z .4
z 4 N1 N1 14 14 N4.4 0 0 0 N0 U0 U --
-0 0 r A
o - a . o
4.' 0 0 l c 0 0 v -
V 'o m
13 14 14r 00 ~ O
~ - 3- -4
NSWC TR 86-108
In these studies, the carbon and Teflon contents were both maintained at 10weight percent. The manufactured electrodes were all found to be rugged andcompletely flexible, although the cathodes containing the 50 percent compressedacetylene black were difficult to process.
Figure 3-48 compares the cycle life performance obtained with the threecarbon types. The cells containing the Shawinigan Acetylene Black were limitedto a maximum depth of discharge of 75 percent which only affected the firstdischarge since the capacity delivered on subsequent cycles was below thisvalue. Equivalent performance was obtained with XC-72R and Shawinigan 50percent compressed carbons while cathodes incorporating Shawinigan 100 percentcompressed carbon delivered somewhat lower capacity over cycle life.
These results show that, at the 10 percent binder level, no significantimprovement in performance is achieved with the Shawinigan Acetylene Blacks asconductive diluents. Therefore, due to its superior processingcharacteristics. Vulcan XC-72R remained the carbon of choice for roll milledcathodes.
Effects of Binder Content 16
Because binders are inert, insulating materials, they can often lower theelectrical conductivity of manufactured electrodes and thus degradeperformance. In general, therefore, it is desirable to maintain the binderconcentration as low as possible, particularly with low conductivity cathodematerials such as V205.
To investigate the effects of binder content on cell performance, rollmilled cathodes were evaluated having Teflon contents of 10, 5. 3. and 1 weightpercent. All electrodes contained 10 weight percent of Vulcan XC-72R carbon asthe conductive diluent. It was found that, even at one percent binder level.the rolled electrodes exhibited excellent mechanical properties.
3-97
NSWC TR 86-108
S00*9t
00,20
C 0
00,01
+ 00
C4 ~ cofL.7 2 00*9Iz E-000z 0
>~ c3 W
14 Cn 44E'C 00 ,0
I E-4
00,0
II W/ I +Oasq
I:Z illoia HOVR -
3-98
NSWC TR 86-108
The cycle results are shown in Figure 3-49. Surprisingly, no significantdifference in cell performance was observed indicating that binder content isnot critical over the range of 1-10 weight percent.
Although no improvement in performance was obtained, the ability tomanufacture rugged, flexible cathodes with binder contents as low as 1 weightpercent is significant. Any reduction in binder or conductive diluent willresult in a corresponding increase in energy density since the cathode willcontain a higher percentage of active material.
Effects of Carbon Content
The effects of carbon content were evaluated using cathodes containing 5,
10. 20 and 30 weight percent XC-72R carbon. In these tests, the Teflon bindercontent was fixed at 5 w-ight percent.
The cycle results are shown in Figure 3-50. Again, the results weresomewhat surprising. With V2 0 having low electrical conductivity, it wasexpected that increasing the cirbon content of the cathodes would bebeneficial. Instead, it was found that the performance dropped significantlywith the higher carbon loadings. In addition, the results show that no loss inperformance occurs when the carbon content is reduced to 5 weight percent whichalso offers benefits with respect to increasing the energy density of V205cathodes.
Effects of V20 5 Particle Size 16
To determine the effects of V 20 particle size on performance. cathodeswere manufactured using Lot CML-V2-803 V205 : As discussed previously, this
material was manufactured from ground ammonium metavanadate in an effort toreduce the particle size of V205. The cathodes contained 10 weight percentVulcan XC-72R carbon and 5 weight percent Teflon binder. Figure 3-51 comparesthe particle size distribution of Lot C-V2-003 with that of Metallurg gradeof V205. As can be seen, there is a distinct difference in thesematerials, particularly at the extremes of the distribution.
Limited cycle life performance results for the two grades of V.,05 arecompared in Figure 3-52. As can be seen, a significant improvement inperformance was achieved with the cell incorporating the fine particle size
V205 indicating that particle size is a key factor affecting performance ofroll milled cathodes.
Summary
The results of these studies have shown that significant improvements in
the performance of roll milled cathodes can be achieved using finer particlesize V20. Therefore, roll milled cathodes have been selected for use inthe prot6type hardware cells to be developed in Phase II of this program.
Final optimization of these cathodes will be conducted in the early stages ofPhase II. encompassing both higher discharge rates and low temperatureoperation. 3-99
NSW.C TR 86-108
Ll002 C)
W GIcc w
E -4
zz 0
u~ 00,9 18,
C Cz
- ~ ~ ~ O 0 00~ ~~j11 11'- .. >.'C
E- Z E. r
00 r.4 C
3--10u000
NSWC TR 86-108
z 7-
00 '1
on on C4
0 021
x+
? z
+ aq-- IOU
U 4 -
3-101
NSWC TR 86-108
PARTICLE SIZE ANALYSIS OF V205
Metaluug Lot 003
Percent of Sample60
50
40
30
20
10 F
f< 46- 5 - - 12-1 >150
Particle Sle in werau
FIGURE 3-51. PARTICLE SIZE DISTRIBUTIONS FOR THE METALLURGV205 VERSUS THAT FOR V20 5 MANUFACTURED FROMGROUND AMMONIUM METAVANADATE (LOT CML-V2-003)
3-102
NSWC TR 86-108
!U 0 ~
4
0
C4 .I CD
0+ 0
.. >4 ..
-± 0 HN
000
Q m ID
(1W3I NO 03V
3-100
NSWC TR 86-108
V2 0 5 PERFORMANCE DEMONSTRATION
Once the cathode technology had been established, additional tests were
conducted to evaluate the performance of the rechargeable Li/V 2O5 system
versus the performance guidelines of the program. In particular, we wanted to
demonstrate that the selected system could indeed achieve 100 cycles withoutsignificant degradation in capacity and to begin to explore the effects ofhigher rates of discharge on extended cycle life performance. We also wantedto verify that CO2 was not detrimental to cell performance since, to this
point. CO had only been evaluated in anode studies and not in completecells. Ail tests were conducted using 2-plate wick cells incorporating roll
milled cathodes composed of 85 w/o V205, 10 w/o Vulcan XC-72R carbon, and
5 w/o Teflon binder.
Figure 3-53 shows the results ior extended cycle life tests at a discharge
rate of 1 mA/cm 2 (approximately C/7 rate). The results for two cells tested
at ambient temperature are shown: one with CO and one without CO2. Forthe first 87 cycles, these two cells performeg similarly, thus demonstratingthat CO has no detrimental effects on cell performance. On the 87th cycle.
the celi without CO2 was inadvertently overdischarged resulting in a rapid
decline in capacity in subsequent cycles. During the 105th cycle, the charge
cutoff was increased from 3.7V to 3.9V which significantly improved the outputalthough the cell never recovered to its original level. The cell containing
CO delivered an average capacity of 0.8 F/mole through 110 cycles afterwhich the capacity declined rapidly. Capacity loss in this cell was attributedto depletion of the anode. The calculated lithium cycling efficiency, based on110 cycles, was 93 percent. identical to the results achieved in the lithiumcycling experiments with the CO -doped solution. The cell tested at -20Cyielded an average capacity of 6.6 F/mole over 160 cycles.
Figure 3-54 shows the extended cycle life results for a cell tested at a
discharge rate ot 5 mA/cm2 (C/1.3 rate). A total of 205 cycles were achieved
at an average capacity of 0.55 F/mole. During the extended testing, this cellwas also inadvertently overdischarged due to an equipment problem. On the 28th
cycle, the cell was discharged to a total depth of 1.82 F/mole, thus
representing a fairly severe overdischarge abuse. Significantly, however, no
substantial loss in performance occurred as a result of the overdischarge.
Table 3-37 lists the average voltages and active material energy densities
realized at various points throughout the cycle life tests described above.
The results of these tests serve to demonstrate the extended cycle life
capabilities of the Li/V 05 system, showing that the 100 cycle deepdischarge goal can indeei 5e achieved with this technology. We have also
demonstrated that long cycle life performance can be achieved at low
temperatures and at moderate rates of discharge, although further improvements
in operating efficiency are needed under these operating conditions. We
believe, however, that these improvements can be readily achieved through
further optimization of the cathode structure.
One of the key concerns with insertion-type cathode materials is their
susceptibility to structural damage during overdischarge resulting in loss of
3-104
NSWC TR 8u~-108
z -1
0j w
Z H:
F *4 u" rr
L - - -
00 I-.
Hn u
cc~ ur
Z0 cn 0
(z -l <1 -4 -.3 >
'J2.1 0 .~ . 0 0
00 =00.- CDE- *C4 0C/C'w
I-4a, - ~ 00I-D :3f.... ; 4
H ' JJ LsZW-a -- 2 D000
H 04
od H 0 O± I9u H- c 'JJ
(0 CUj
(nqow/d T NO G3SV9)IN3Sd :ANIIA 39HVHDSIUI
3-105
NStC TR 86-108
a..
. 000
CN
Q00 09 I
z 0 E
,a 80001,,
C -co ,z" .
H >
00 40
(-' 00
- >-; 8 "9
-44-II '.. r.1CD
-00,00
3-0
E -
Ul2+ Qa
IN N 0 3S q
ED21IJ >9H DSIG
3-1060.
NSWC TR 86-108
TABLE 3-37. PERFORMANCE SUMMARY FOR Li/V205 LABORATORY CELLS
TESTED UNDER EXTENDED CYCLE LIFE CONDITIONS
Delivered*Discharge Test Delivered Average Energy
Cycle Rate Temperature Capacity. Voltage. Density.No. mA/cm2 0C F/Mole V Wh/Kg Notes
1 1 Ambient (20-24°C) 0.89 3.23 410 Without CO22 1 " 0.75 3.21 347 (Cell B in
The above energy densities are based on active materials only.
reversibility. Although we have not yet done a systematic investigation, theinadvertent overdischarges that occurred during our extended cycle life testsindicate that the Li/V O5 technology can sustain significant overdischargebeyond I F/moLe withoui catastrophic loss of performance. This area will bemore fully evaluated in Phase II so as to establish appropriate operatingboundaries for this system.
3-107
NSWC TR 86-108
CHAPTER 4
CONCLUSIONS
At the onset of this program, the room-temperature lithium rechargeabletechnology faced performance limitations with respect to rate capability andlow temperature operability. To overcome these performance deficiencies, thisprogram undertook the development of four high energy density cathode materials(V 0 TiS-, V S and LiCoO ) in conjunction with the use of ester-basedelczrolyt soiutions becausi of their superior solution conductivity.
An electrolyte formulation composed of 2M LiAsF + 0.4 LiBF /methylformate saturated with CO was successfully developed with the followingdemonstrated capabilities:
o Solution Conductivity--43 mmh^ cm at room temperature and 13 mmho/cmat -40 C. This solution at -40 C is still three times moreconductive than the more commonly used ether-based solution at ambienttemperature.
o 93% Lithium Efficiency--this lithium efficiency enables the use ofmethyl formate-based solution in practical hardware. In the case ofLi/V 205 cell, cyclability up to 50 cycles can be projected at the100% depth of discharge based on Li/V2 05 ratio of 3:1.
Of the four cathode materials studied (V 0 5 , TiS 2 , V2 S5, and LiCoO1 ),V205 emerged to be the best overall performei:
o Cyclability surpassing 100 cycles--cycled cell capacity can be held ata level corresponding to 80 percent cathode discharge efficiencythroughout the life of the cell before reaching the lithium depletionpoint. To a large extent, the ability to achieve this high level ofcathode utilization is due to our focused development of the cathodeprocessing technology. Particle size of the V205 was found to bekey to enhanced cell performance and electrode integrity was criticalto achieving cyclability without degrading cell capacity.
Li/V 0 laboratory cell produces 366 Wh/kg (or 166 Wh/lb) at the100ti ;ycle when tested at room temperature. Using this value.practical energy densities of 50 Wh/lb and 60 Wh/lb can be projectedfor "D" cell and No. 6 size cell, respectively.
4-1
NSWC TR 86-108
0 Discharge current density surpassing 5 mA/cm2 __ti allows
addressing application of the Li/V 0 technology in missionsrequiring rate capability up to CD
The other three cathode materials were dropped from further
development because of their general incompatibility with ester-based
solutions. Specifically. severe cointercalation was observed between
TiS 2 and methyl formate which adversely affect cycle performance.Discharge capacity was severely degraded even on the first cycle for
V S~ cat hode in methyl formate-based solution. Finally, althoughLl 2was demonstrated to be reversible and offered to be a high
cell potential cathode. the high cell potential needed to charge this
cathode system decomposed the solvent during charge.
In closing. Phase I of this program successfully selected a rechargeable
technology (Li/V 2 o5 ) that is ready for transitioning to a more hardwaredevelopment phase. Phase I demonstrated overall performance capabilities of
the Li/V 20 technology at the laboratory cell level, and the objective of
Phase II i; to demonstrate the stated capabilities of Li/V 20 5 cell at the
hardware level via a 30 Ah cell.
4-2
NSWC TR 86-108
REFERENCES
1. Ebner. W.B. and Lin, H.W., "High Cycle Life Rechargeable Lithium - OrganicElectrolyte Cell." Third Quarterly Report. Contract DAAK20-84-C-0412(LABCOM), Honeywell Inc., Oct 1985.
2. Winn, D.A., "Titanium Disulphide: A Solid Solution Electrode," Mat. Res.Bull., 11. 1976, p. 559.
3. Yamamoto, T., Kikkawa, S. and Koizumi, M., "Effect of Nonstoichiometry andSolvent on Discharge Property of Li/TiS2 Battery." J. Electrochem. Soc.,131, 1984. p. 1343.
4. Rao, B.M.L. and Shropshire. J.A.. "Effect of Sulfur Impurities on Li/TiS2Cells," J. Electrochem. Soc., 128. 1981, p. 942.
5. Whittingham, M.S. and Panella, J.A., "Formation of Stoichiometric TitaniumDisulfide," Mat. Res. Bull., 16. 1981, p. 37.
6. Mizushima, K., Jones, P.C., Wiseman, P.J., and Goodenough, J.B.,"Lix Coo (0 < X < 1): A New Cathode Material for Batteries of HighEnergy Density." Mat. Res. Bull., 15, 1980, p. 783.
7. Deshpande, S.L. and Bennion, D.N., "Lithium Dimethyl Sulfite GraphiteCell," J. Electrochem. Soc., 125, 1978, p. 687.
8. Ebner, W.B. and Walk, C.R.. "LiAsF - Methyl Formate ElectrolyteSolutions," Proceedings of the 27tg Power Sources Symposium, 21-26 Jun1976, p. 48.
9. Honeywell Internal Development Program.
10. Jacobson, A.J. and Rich, S.M., "Electrochemistry of Amorphous V2 S5 inLithium Cells," J. Electrochem. Soc., 147. 1980, p. 779.
11. Lin, H.W. and Ebner, W.B., "High Cycle Life Rechargeable Lithium - OrganicElectrolyte Cells." Second Quarterly Report, DAAK20-84-C-0412 (LABCOM).Honeywell Inc.. Aug 1985.
5-1
NSWC TR 86-108
REFERENCES (Cont.)
12. Murphy. D.N. Christian. P.A., DiSalvo. F.J. and Waszczak. J.V.. "LithiumIncorporation by Vanadium Pentoxide." Inorganic Chemistry, 18, 1979.p. 2800.
13. Foos, J.S. and Rembetsy, L.M., "Lithium Cycling in Sulfolane-BasedElectrolytes," Extended Abstracts. 83-2, Electrochem. Soc., Fall Meeting.Washington, D.C., 9-14 Oct 1983, p. 117.
14. Glugla, P.G.. "Lithium Cycling Behavior in 2-Methyltetrahydrofuran WithAlcohol Additives," J. Electrochem. Soc., 130. 1983. p. 113.
15. Abraham, K.M.. Foos, J.S. and Goldman. J.L.. "Long Cycle-Life SecondaryLithium Cells Utilizing Tetrahydrofuran." J. Electrochem. Soc., 131. 1984.p. 2197.
16. Ebner, W.B. and Lin, H.W.. "High Cycle Life Rechargeable Lithium - OrganicElectrolyte Cells." Fourth Quarterly Report. Contract DAAK20-84-C-0412(LABCOM). Honeywell Inc., Feb 1986.
5-2
NSWC TR 86-108
DISTRIBUTION
Copies Co pies
Defense Technical Information Center Naval Civil Engineering Lab
Cameron Station Attn: Alan Inouye (Code L43)
Alexandria, VA 22314 12 Port Hueneme. CA 93043-5003
Utfice of Naval Technology Commander.
Atnt: Code 232 Naval Weapons Support Center(Dr. A. J. Faulstich) 2 Attn: J. Marusek (Code 30512)
80.) N. Quincy Street Crane, IN 47522Arlijngton, VA 22217-5000
Naval Underwater Systems Center
Institute for Defense Analyses Attn: J. Moden (Code SB332)R&E Support Division Newport, RI 02840
400 Army-Navy DriveArlington, VA 22202 1 David W. Taylor Naval Ship
R&D Center
Commander Attn: W.J. Levendahl (Code 2703)
Space and Naval Warfare Systems Annapolis LaboratoryCommand Annapolis, MD 21402
Washington, DC 20363-5100 1Air Force of Scientific Research
Office of Naval Research Attn: R.A. OsteryoungAttn: Code 1113 (Dr. R.J. Nowak) 2 Directorate of Chemical Science800 N. Quincey Street 1400 Wilson BoulevardArlington. VA 22217-5000 Arlington, VA 22209
Naval Ocean Systems Center Air Force Aero Propulsion Laboratory
San Diego. CA 92152 Wright-Patterson AFB, OH 45433
Naval Coastal Systems Center Office of Chief of Research and
Attn: R. Johnson (Code 410T) 1 DevelopmentT. Sherman (Code 4210) 1 Department of the Army
Panama City, FL 32407 Attn: Dr. S.J. MagramEnergy Conversion Branch
Mare Island Naval Shipyard Room 410. Highland BuildingAttn: S.G. Marsh (Code 280.39/0204) 1 Washington, DC 20315Ocean Engineering Department
Vallejo, CA 94592 U.S. Army Research Office
Attn: B.F. SpielvogelP.O. Box 12211Research Triangle Park, NC 27709
(1)
i " -- • m.- mmmm m m m
NSWC TR 86-108
U.S. Army Electronics Command California Institute of TechnologyAttn: Dr. M. Salomon (Code SLCET-PR) 1 Attn: Dr. R. Somoano
Dr. W.K. Behl (Code SLCET-PR) 1 Dr. H. FrankDr. M. Binder (Code SLCET-PR) 1 G. Halpert
Fort Monmouth, NJ 07703 Jet 1ropulsion Laboratory4800 Oak Grove Drive
Army Material and Mechanical Research Pasadena, CA 91103Center
Attn: J.J. DeMarco 1 Sandia LaboratoriesWatertown, MA 02172 Attn: Dr. S. Levy
Albuquerque. NM 87115USA Mobility Equipment R&D CommandAttn: J. Sullivan (Code DRXFB) 1 ESB Research Center
(Code DRME-EC) 1 Attn: LibraryElectrochemical Division 19 W. College AvenueFort Belvoir, VA 22060 Yardley, PA 19067
Department of Energy EIC-CorporationAttn: Dr. A. Landgrebe ATtn: Dr. K.M. Abraham
(Code MS E-463) 1 111 Downey StreetEnergy Research and Development Norwood, MA 02062
AgencyDivision of Applied Technology Boeing Aerospace CompanyWashington. DC 20545 Attn: S. Gross 1
C.J. Johnson 1NASA Headquarters P.O. Box 3999Attn: Dr. J. H. Ambrus 1 Seattle, WA 98124
(Code RTS-6) IWashington. DC 20546 Moli Energy Ltd.
Attn: Dr. J.A. StilesCentral Intelliger-e Agency 3958 Myrtle StreetAttn: Dr. T.X. Mahy 3 Burnaby, B.C. Canada V5C 4G2c/o OTSWashington. DC 20505 Library of Congress
ALtn: Gift and Exchange Div. 4NASA Goddard Space Flight Center Washington, DC 20540
Attn: T. Hennigan (Code 716.2) 1Greenbelt, MD 20771
Bell LaboratoriesAttn: Dr. J.J. Auborn I
Dr. J. Broadhead 1600 Mountain AvenueMurray Hill, NJ 07974
(2)
NSWC TR 86-108
Eagle-Picher Industries, Inc. Duracell Int., Inc.
Attn: D.R. Cottingham 1 Attn: Dr. A.N. Dey
J.Dines 1 Duracell Research Center
Electronics Division, Couples 37 A St.Department Needham. MA 02194
P.O. Box 47Joplin, MO 64801 Power Conversion. Inc.
70 MacQuesten Parkway S.
Foote Mineral Company Mount Vernon. NYH 10550
Attn: H.R. Grady 1Exton. PA 19341 Union Carbide Battery Products
Defense Systems GroupPower Sources Center Battery Engineering Inc.
104 Rock Road Attn: Dr. N. Marincic
Horsham, PA 19044 Dr. C.R. Schlaikjer1636 Hyde Park Avenue
Lockheed Missiles and Space Hyde Park, MA 02136
Company. Inc.Attn: Library 1 Eveready Battery Co.
Lockheed Palo Alto Research Attn: Dr. G.E. Blomgren
Laboratory 25225 Detroit Road
3251 Hanover Street P.O. Box 45035
Palo Alto, CA 94304 Westlake, OH 44145
Ballard Research Inc.
Attn: Dr. A.C. Harkness 1
1164 15 St. WNorth Vancouver, B.C.
Canada V7P 1M9
Duracell Int., Inc.
Attn: B. McDonald 1
Battery DivisionSouth BroadwayTarrytown, NY 10591
(3)
NSWC TR 86-108
RAY-O-VAC Electrochimica CorporationAttn: R. Foster Udell 1 Attn: Dr. M. Eisenberg.dl East Washington Avenue 20 Kelly CourtMadison. WI 53703 Menlo Park. CA 94025
TRW Systems Electrochemistry ConsultantsAttn: Ed Moon. Rm. 2251 1 Attn: Dr. B.B. OwensBldg. 0-1 4707 Lyndale Avenue NorthOne Space Park Minneapolis. MN 55430Redondo Beach, CA 90278
Electrochem Industries
ALTUS Corporation Attn: Dr. W.D.K. Clark
Attn: G.L. Griffin 1 10.000 Wehrle Drive1610 Crane Court Clarence. NY 14031San Jose. CA 95112
